>> Announcer: From around the globe, it's theCUBE with digital coverage of AWS Public Sector Partner Awards. Brought to you by Amazon Web Services. >> Hi friend, welcome to this CUBE coverage of AWS Public Sector Partner Awards Program. I'm John Furrier your host of theCUBE. We've two great guests here, Travis Hartman Director of Analytics and Weather at Maxar Technologies, and Vijay Tallapragada who's the Chief Modeling and Data Assimulation Branch at NOAH. Tell us about the success of this. What's the big deal? Take us through the award and why Maxar. What do you guys do? >> Yeah, so Maxar is an organization that does a lot of different activities in earth intelligence as well as space. We have about 4,000 employees around the world. One side of the economy works on space infrastructure actually building satellites, and all the infrastructure that's going to help get us back to the moon, and things like that, and then on the other side we have an earth intelligence group which is where I sit, and we leverage remote sensing information, earth science information to help people better understand how and what they do might impact the earth, or how the earth, in its activities, might impact their business mission or operations. So what we wanted to set out to do is help people better understand how weather could impact their mission, businesses, or operations. A big element of that was doing it with speed. So we knew NOAH had capabilities of running numerical weather prediction models and very traditional on-prem, big, beefy, high performance supercomputers, but we wanted to do it in the cloud. We wanted to use AWS as a key partner. So we collaborated with Vijay and NOAH and his teams there to help pull that off. They gave us access, public domain information but they showed us the right places to look. We've had some of our research scientists talkin' and yeah, after a pretty short effort, it didn't take a lot of time, we were able to pull something off a lot of people didn't think was possible. And we got pretty excited once we saw some of the outcomes. >> Travis, Vijay was just mentioning the relationship. Can you talk about the relationship together? Because this is not your classic Amazon Partner client relationship that you have. You guys have been partnering together, Vijay and your team, with AWS. Talk about the relationship and how Amazon played because it's a unique partnership. Explain in more detail, that specific relationship. >> Yeah, with Maxar and AWS, our partnership has gone back a number of years. Maxar being a fairly large organization, there's lots of different activities. I think Maxar was the first client of AWS Snowmobile where they had the big tractor trailer backed up to a data center, load all the data in, and then take it to an AWS data center. We were the first users of that 'cause we had over a hundred petabytes of satellite imagery in an archive that just movin' it across the internet it'd probably still be goin'. So the Snowmobile was a good success story for us but just with the amount of data that we have, the amount of data we collect every day, and all the analytics that we're running on it, whether it's in an HPC environment or the scalable AIML, we're able to scale out that architecture, scale out the compute, the much easier dynamic and really cost-effective way with AWS 'cause when we don't need to use the machines, we turn 'em off. We don't have a big data center sittin' somewhere where we have to have security, have all the overhead costs of just keeping the lights on, literally. AWS allows us to run our organization in a much more efficient way. And NOAH, they're seeing some of that same success story that we're seeing, as far as how they could use the cloud for accelerating research, accelerating how the advancement of numerical weather prediction from the United States can benefit from cloud, from cloud architecture, cloud compute, and things like that. And I think a lot of the stuff that we've done here at Maxar, with our HPC solution in the cloud is something that's pretty interesting to NOAH and it's a good opportunity for us to continue our collaboration. >> If I could drill down on that solution architecture for a minute, how did you guys set up the services and what lessons did you learn from that process? >> We're still learnin' is probably the short answer, but it all started with our people. We have some really strong engineers, really strong data scientists that fundamentally have a background in meteorology or atmospheric science, so they understand the physics of, you know, why the wind blows the way it does and why clouds do what clouds do. But we also, having a key strategic partnership with AWS, we were able to tap into some of their subject-matter experts, and we really put those people together and come up with new solutions and new, innovative ideas, stuff that people hadn't tried before. We were able to steer a little bit of AWS's product roadmap as far as what we were tryin' to do and how their current technology might not have been able to support it, but by interacting with us, gave them some ideas as far as what the tech had to move towards, and then that's what allowed us to move in a pretty quick fashion. It's neat stuff, technology, but it really comes down to the people. I feel very honored and privileged to work with both great people here, at Maxar, as well as AWS, as well as bein' able to collaborate with the great teams at NOAH. It's been a lot of fun. >> Well Travis, got a great example, I think it's a template that can be applied to many other areas, certainly even beyond. You've got a large scale, multi-scale situation, there. Congratulations. Final question, what does it mean to be an award winner for AWS Partner Awards? As part of the show, you're the best-in-show for HPC. What's it like? What's the feeling? Give is a quick stub from the field. >> Yeah, I mean, I don't know if there's really a lot of good words that can kind of sum it up. I shared the news with the team last night and you know, there were a lot of, lot of good responses that came from it. A lot of people think it's cool, and at the end of the day, a lot of people on our team took a hobby or a passion of weather and turned it into a career. And being acknowledged and recognized by groups like AWS for best solution in a particular thing, I think we take a lot of that to heart and we're very honored and proud of what we're able to do and proud that other people recognize the neat stuff that we're doin'. >> Well, certainly takin' advantage of the cloud which is large scale, but you're on a great wave, you've got a great area. I mean, weather, you talk about weather, it's exciting, dynamic, it's always changing, it's big data, it's large scale. So you got a lot of problems to solve and a lot of impact too, when you get it right. So congratulations on an excellent-- >> Thank you very much. >> Great mission. >> Thank you. >> Love what you do, love to followup again and maybe do another interview, and talk about the impact of weather and all the HPC kind of down the road. Travis, thank you very much. >> Thank you, appreciate it. >> Good to see you. >> Thank you, glad to be here. >> So NOAH, National Oceanic Atmospheric Administration, National Weather Center, National Center for Environmental Predictions, Environmental Modeling Center, that's your organization. You guys are competing to be the best in the world. Tell us what you guys do at a high level, then we'll jump into some of the successes. >> So the National Weather Service is responsible for providing weather forecasts to save lives and property, and improve the economy of the nation. And as part of that, the National Weather Service is responsible for providing data and also the forecast to the public and to the industry. We are responsible for providing the guidance on how they create the forecasts. So we are, at the Environmental Modeling Center, the nation's finest institute in advancing our numerical weather prediction modeling, government, and a nucleation of all the data that's available from the world to initialize our models and provide the future state of the atmosphere from hours all the way to seasons and years. And that's the kind of the range of products that we download and provide. Our key for managing the emergency of services and hazard management and mitigation, and also improve in the nation's economy by preparing well in advance, for the future events. And it's a science-based organization and we have world-class scientists working in this organization. I manage about 170 of them at the Environmental Modeling Center. They're all PhDs from various disciplines, mostly from meteorology, atmospheric sciences, oceanography, land surface modeling, space weather, all weather-related areas, and the mathematics and computer science. And we are at the stage where we are probably the most doubled up, advanced modeling center that we use almost all possible computational services available in the world, so this is heavily computational in terms of use of data, use of computers, use of all the power that we can get, and we have a 3.5 protoflop machine that we use to provide these weather forecasts. And they provide these services every hour for some census like we see the weather outbreaks and for every three hours for hurricanes, and for every six hours for the regular weather like precipitation, the temperature forecasts. So all the data that you see coming out from either the public media or the government agencies, they all are originated in our center and disseminated in various forms. And I think NOAH is the only center in the world that provides all this information free of cost. So it is a public service organization and we pride in our service to the society. >> Well, I love your title, Chief Modeling and Data Assimulation title, branch over all these organizations. This is, weather's critical. I want to get your thoughts 'cause we were talking before you came on about how the hurricane Katrina was something that really kind of forced everyone to kind of rethink things. Weather is an evolving system so it's always changing. Either there's a catastrophe or something happens, or you're trying to be proactive, predicting say, whether it's a fire season in California, all kinds of things goin' on. It's always hard to get a certain prediction. You have big jobs, there's a lot of data, you need horsepower, you need computing, you need to stand up some HPC. Take us through the thinking around the organization and what's the impact that you see, because weather does have that impact. >> So traditionally, you know, as you mentioned there are various weather phenomena that you described like the fiber of the hurricanes, the heavy precipitation, the flooding, so we download solutions for individual weather phenomena. And we have grown in that direction by downloading separate solutions for separate problems. And very soon, it became obvious that we cannot manage all these independent modeling systems to provide the best possible forecasts. So the thinking had to be changed. And then there is another bigger problem is that there's a lot of research going out in the community, like the academic institutes, the universities, other government labs. There are several people working in these areas and all their work is not necessarily a coordinated government act duty, that we cannot take advantage, and there are no incentives for people to come and contribute towards the mission that we are engaged in. So that actually prompted to change the direction of thinking, and as you mentioned, hurricane Katrina was an eye-opener. We have the best forecasts, but the dissemination of that information was not probably accurate enough, and also there is a lot of room for improvement in predicting these catastrophic events. >> How are you guys using AWS? Because HPC, high performance computing, I mean, you can't ask for more resources than the massive cloud that is Amazon. How has that helped you? Can you take a minute to explain, walk us through AWS partnership? >> There are a few examples I can cite, but before then, I would really like to appreciate Travis Hartman from Maxar who is probably the only private sector partner that we had in the beginning. And now, we are expanding on that. So we were able to share our immunity cords with Maxar and with our help, they were able to establish this entire modeling system as it is done in operations at NOAH. They were able to reproduce our operational forecasts using the cloud resources and then they went ahead and did even more by scaling the modeling systems as they can run even faster and quicker than what NOAH operations can do. So that gives you one example of how the cloud can be used. You know, the same forecast that we produce globally, which will take about eight minutes per day, and Maxar was able to do it much faster, like 50% improvement in the efficiency of the cords. And now, the one advantage of this is that the improvements that Maxar or other collaborators are using our cords, that they're putting into the system, are coming back to us. So we take advantage of that in improving the efficiency in operations. So this like a win-win situation for both of what part is fitting in the R&D and what using in operations. And on top of it, you can create multiple conflagrations of this model in various instances on the cloud where you can run it more efficiently and you can create an ensemble of solutions that can be catered to individual needs. And the one additional thing I wanted to mention about the user cloud is that this is like when you have a need, you can surge the compute, you can instantiate thousands of simulations to test a new innovation, for instance. You don't need to wait for the resources to be done in sequential manner. Instead, you can ramp up the production of these equipments in no time, and without worrying about, of course, the cost is a factor that we need to worry about, but otherwise the capacity is there, the facilities are there to take advantage of the cloud solutions. >> Well Vijay, I'm very impressed with your organization. I'd love to do a followup with you. I love the impact that you're doing. Certainly, the weather impacts society from forecasting disasters and giving people the ability to look at supply chain, whether it's planning for potentially a fire season or a water shortage, or anything goin' on, there. But also it's a template. You are succeeding a new kind of way to innovate with community, with large scale, multi-scale data points, so congratulations. >> Thank you. >> Thank you very much. I'm John Furrier here, part of AWS Partner Awards Program, best HPC solution. Great example, great use case, great conversation. Thanks for watching. Two great interviews here, as part of AWS Public Sector Partner Awards Program. I'm John Furrier. The best-in-show for HPC solutions, Travis Hartman, Maxar Technologies, and Vijay Tallapragada at NOAH, two great guests. Thanks for watching. (soft electronic music)

>>from around the globe. It's >>the Cube with digital coverage of >>AWS Public Sector Partner Awards >>brought to you by >>Amazon Web services. Everyone, welcome to this cube coverage of AWS Public Sector Partner Awards program. I'm John Furrow, your host of the Cube with two great guests here. Travis Department director of analytics and Weather at Max. Our technologies and VJ teleplay Gotta Who's the chief? Modeling and data a simulation branch at Noah. Tell us about the success of this. What's the big deal? Take us through the award and why Max are what you guys do. >>Yeah, so Macs are is an organization. Does a lot of different activities unearth intelligence as well as space? We have about 4000 employees around the world. One side of the economy works on space infrastructure, actually building satellites on all the infrastructure that's going to help us get us back to the moon and things like that. And then on the other side we have a north of intelligence group, which is where, I said, and we leverage remote sensing information for science information to help people better understand how, how and what they do might impact the Earth or have the earth, and it's activities might impact their business mission. Our operation. So what we wanted to set out to do was help people better understand how weather could impact their mission, business or operations. And a big element of that was doing it with speed. Ah, so we we knew? No. I had capabilities running America weather prediction models and very traditional on Prem. Big, beefy ah, high performance compute supercomputers. But we wanted to do it in The cloud we want to do is AWS is a key part. So we collaborated with B. J and Noah and his team is there to help pull that off. They gave this access public domain information, but they showed us the right places to look. We've had some of the research scientists talking, and after pretty short effort, it didn't take a lot of time. We were able to pull something off that a lot of people didn't think was possible. I'm we got pretty excited. Once we saw some of the outcome >>Travis to be, Jay was just mentioning the relationship. Can you talk about the relationship together because this is not your classic Amazon partner client relationship that you have. You guys have been partnering together V. J and your team with AWS. Talk about the relationship and that and how Amazon plays because it's a unique partnership plane in more detail at specific relationship. >>Yeah, with Max or in AWS. You know, our partnership has gone back A number of years on Macs are being a fairly large organization. There's lots of different activities. I think Max Star was the first client of AWS Snowmobile, where they have the big tractor trailer back up to a data center, load all the data in and then take it to an AWS data center. We were the first users of that because we had over 100 petabytes of satellite imagery and archive that just moving across the Internet would probably still be going. Um, so the snowmobile is a good success story for us, but just with >>the >>amount of data that we have, the amount of data we collect every day and all the analytics that we're running on it, whether it's in an HPC environment or, you know, the scalable Ai ml were able to scale out that architecture scale out that compute the much easier, dynamic and really cost effective way with AWS, because when we don't need to use the machines, we turn them off. We don't have a big data center sitting somewhere. We have to have security, have all the overhead costs of just keeping the lights on. Literally. AWS allows us to run our organization and a much more efficient way. Um and Noah, you know, they're They're seeing some of that same success story that we're seeing as far as how they can use the cloud for accelerating research, accelerating how the advancement of numerical weather prediction from the United States can benefit from cloud from cloud architecture, cloud computer, things like that. And I think a lot of the stuff that we've done here, Max our with our HPC HPC solution in the cloud. It's something that's pretty interesting to know, and it's it's a good opportunity for us to continue our collaboration. >>If I could drill down on that solution architecture for a minute. How did you guys set up the services, and what lessons did you learn from that process? >>We're still learning. It was probably the the short answer, but it all started with our people. Uh, you know, we have some really strong engineers, really strong data scientists that fundamentally have a background in meteorology or atmospheric science, you know? So they understand the physics. So you know why the wind blows is the way it doesn't. Why Cloud's doing clouds to do, Um, but we also having a key strategic partnership with AWS. We really have to tap into some of their subject matter experts. And we really put those people together, you know, and come up with new solutions, new innovative ideas, stuff that people hadn't tried before. We're able to steer a little bit of AWS is product roadmap for is what we were trying to do and how their current technology might not have been able to support it. But by interacting with us gave them some ideas as far as what the tech had to move towards. And then that's that's what allowed us to move pretty quick fashion. Um, you know, it's it's neat stuff technology, but it really comes down to the people. Um, and I feel very honored and privileged to work with both great people here. Attacks are as well as aws, um, as well as being able to collaborate with your great teams. That power, it's been a lot of fun. Well, >>Travis gonna create example? I think it's a template that could be applied to many other areas, certainly even beyond. You've got large scale, multi scale situation there. Congratulations. Final question. What does it mean to be an award winner for AWS Partner Awards as part of the show? You're the best in show for HPC. What's it like? What's the feeling? Give us a quick side from the field? >>Yeah. I mean, I don't know if there's really a lot of good words that kind of sum it up. It's Ah, I shared the news with the team last night, and you know, there are a lot of a lot of good responses that came from a lot of people think it's cool. And at the end of the day, a lot of people on our team, you know, took a hobby or a passion of weather and turned it into a career. Ah, and being acknowledged and recognized by groups like AWS for best solution in a particular thing. Um, I think we take a lot of that to heart. And, ah, we're very honored and proud of what we were able to do and proud that other people recognize the need stuff that we're doing well, >>Certainly taking advantage. The cloud, which is large scale. But you you're on a great wave. You've got a great area. I mean, whether you talk about whether it's exciting, it's dynamic. It's always changing. It's big data. It's large scale. So you get a lot of problems to solve in a lot of impact to get it right. So congratulations on ECs. >>Thank you very much. Great mission. Thank you. >>Love what you do love to follow up again. Maybe do another interview and talk about the impact of weather and all the HPC kind of down the road. But, Travis, thank you very much. >>Thank you. Appreciate it. >>Good to see you. >>Thank you. Good to be here. >>So Noah, National Oceanic Atmospheric Administration, National Weather Center, National Center for Environmental Predictions, Environmental Modeling Center year. That's your organization? You guys are competing to be best in the world. Tell us what you guys do at a high level. Then we'll jump into some of the successes. >>So the national Weather Service is responsible for providing weather forecast to save lives and property and improve the economy of the nation. And that's part of that. That the national weather services responsible for providing data and also the forecasts to the public and the industry and be responsible for providing the guidance on how they create the forecasts. So we are at the Environmental Modeling Center, uh, the nation's finest institute in advancing our numerical weather prediction modelling development, and you play it off all the data that's available from the world to initialize our models and provide the future state of the atmosphere from hours all the way to seasons and years. That's that's the kind of a range of products that we don't lock and provide are our key for managing the emergency services and patch it management and mitigation and also improving the nation's economy by preparing well in advance for the future events. And it's it's a science based organization, and we have ah well class scientists working in this organization. I manage about 170 of them at the moment of modeling center. They're all PhDs from various disciplines, mostly from meteorology, atmospheric sciences, oceanography, land surface modelling space weather, all weather related areas and the mathematics and computer science. And we are at the stage where we are probably the most. Uh huh. Most developed, uh, advanced modelling center that we use almost all possible computational resources available in the world. So this is a really computational in terms of user data, user computer seems off. Uh, all the power that we can get and we have a 3.5 petaflop machine that we use to provide these weather forecasts, and they provide the services every hour. For some sense is like the CDO rather our rates for every three hours for hurricanes and for every six hours for the regular, Rather like the participation, uh, the temperature forecast. So all the data that you see coming out from either the public media, our department agencies, they are originated in our center and disseminated in various forms. I think no one is the only center in the world that provides all this information for your past. So it is, ah, public service organization and we riding on a visa with society. >>We'll I love your title, Chief modeling and data, a simulation title branch of a lot of these organizations. This >>is >>whether it's ever critical. I want to get your thoughts cause we were talking before we came on about how the Hurricane Katrina was something that really kind of forcing you to rethink things. Whether it is an evolving system, it's always changing. Either the catastrophe or something happens. Were you trying to proactive predicting, say, whether it's a fire season in California, all kinds of things going on that's not It's always hard to get a certain prediction. You have big job. It's a lot of data you need. Horsepower need computing. You need to stand up. Some HPC take us through like like the thinking around the organization. And what was The impact is that you see, because whether does have that impact. >>So traditionally, you know, as you mentioned, there are radius weather phenomenon that you describe like the five rather the Americans, every presentation, the flooding. So we developed solutions for individual weather phenomena, and, uh, we have grown in that direction by developing separate solutions for separate problems. And very soon it became obvious that we cannot manage all these independent modeling systems to provide the best possible forecasts. So the thinking has to be changed. And then there is Another big problem is that there's a lot of research going out in the community like the academic institutes, the universities, other government labs. There are several people working in these areas, and all their work is not necessarily a coordinated, uh, development activity that we cannot take advantage. And they have no incentive for people to come and contribute towards the mission that we are engaged in. So that actually prompted to change the direction of thinking. And as you mentioned, Hurricane Katrina was an eye opener. We had the best forecasts, but the dissemination of that information waas not probably accurate enough, and also there is a lot of room for improvement in predicting these catastrophic events. How are >>you guys using AWS? Because HPC high performance computing I mean you can't ask for more resources in the massive cloud that is Amazon. How is that help to you? Can you take a minute to explain, but walk us through? >>What? >>Aws? There >>are a few example. Second site. But before then, I would like to really appreciate a Travis Hartman from Max. Are you know who is probably the only private sector partner that we had in the beginning. And now we're expanding on. That s so we were able to share our community. Cores with Max are and without how they were able to establish this and drive modeling system as it is done in operations that Noah and they were able to reproduce operational forecast using the cloud resources. And then they went ahead and did even more by scaling the modeling systems is that it can run even faster and quicker them are what insert no operations can do. So that gives us one example of how the cloud can be used. You know, the same forecast that we produce, ah, globally, which will take about eight minutes per day. And, uh, Max I was able to do it much faster, like 50% improvement and in the efficiency of the colors. And now the one piece of this is that the improvements that matter are other collaborators are using, or cords that they're putting into the system are coming back to us. So we take advantage of that, improving the efficiency in operations. So this is that this is like a win win situation for both, uh, who are participating in the R and D on who are using it in operations, and on top of it, you can create multiple configurations of this model in various instances on the cloud when you can run it more efficiently and you can create an ensemble of solutions that can be captured toe individual needs. And the one additional thing I want to mention about User Cloud is, is that you know, this is like when you have a need, you can search the compute you can. Instead she 8000 sub simulations to test a new innovation. For instance, you don't need to wait for the resources to be done in a sequential manner. Instead, you can ramp up the production off these apartments in no kind and without Don't worry about. Of course, the cost is the fact that we need to worry about, but otherwise the capacity is there. The facilities are reacting to take advantage of the cloud solutions. If I'm a >>computer scientist person, I'm working on a project. Now I have all this goodness in the cloud, how's morale been and what's the reaction been like from from people doing the work. Because usually the bottleneck has been like I gotta provision resource. I gotta send a procurement request for some servers or I want to really push some load. And right now, I got a critical juncture. I mean, it's got a push morale up a bit, and you talk about the impact to the psychology of the people in your organization. >>Um, I haven't. I have two answers to this question. One from a scientist perspective like me. You know, I was not a computer scientist from the beginning, but I became a software engineer, kind of because I have to work with these software and hardware stuff more more on solving the computational problems than the critical problems. So people like us who have invested their careers in improving the science, they were not care whether it's ah, uh hbc on premise Cloud, what will be delighted to have, uh, resources available alleviate that they can drive. But on the other hand, the computer computational engineers are software engineers who are entering into this field. I think they are probably the most excited because of these emerging opportunities. And so there is a kind of a friction between the scientific and the computational aspects off personnel, I would say. But that difference is slowly raising on and we are working together as never before. So the collective moral is very high to take advantage of these resources and opportunities. I think way of making the we're going in the right direction. >>It's so much faster. I mean, in the old days, you write a paper, you got to get some traction. Gonna do a pilot now It's like you run an experiment, get it out there. VJ I'm very impressed with the organization. Love to do a follow up with you. I love the impact that you're doing certainly in the weather impact society from forecasting disasters and giving people the ability to look at supply chain, whether it's providing for potentially a fire season or water shortage or anything going on there. But also it's a template. You're exceeding a new kind of waiting to innovate with community with large scale, multi scale data points. So congratulations and >>thank you. >>Thank you very much. I'm John Furrier here part of AWS partner Awards program. Best HPC solution. Great. Great Example. Great use case. Great conversation. Thanks for watching two great interviews. Here is part of AWS Public Sector Partner Awards program. I'm John Furrier. The best in show for HPC Solutions. China's Hartman Max, our technologies and Vijay tell Apartado at Noah. Two great guests. Thanks for watching. Yeah, Yeah, yeah, yeah, yeah, yeah

>> Narrator: From theCUBE studios in Palo Alto in Boston, connecting with thought leaders all around the world, this is a cube conversation. >> Everyone welcome to this cube conversation here in the Palo Alto cube studios. I'm John Furrier your host of theCUBE, with a great guest here, Shanthi Vigneshwaran, who is with the Office of Strategic programs in the Center for Drug Evaluation and Research within the US Food and Drug Administration, FDA, is the Informatica Intelligent Disrupter of the Year award. Congratulations, Shanthi welcome to this cube conversation. Thanks for joining me. >> Thank you for having me. >> Congratulations on being the Informatica Intelligent Disrupter of the year award. Tell us more about the organization. I see FDA everyone's probably concerned these days making sure things going faster and faster, more complex, more things are happening. Tell us about your organization and what you work on. >> FDA is huge, our organization is Center for Drug Evaluation research. And its core mission is to promote public health by ensuring the availability of safety and effective drugs. For example any drugs you go and buy it in the pharmacy today, Our administration helps in trying to approve them and make sure it's so in term of quality and integrity of the marketed products in the industry. My office is specifically Office of strategic programs whose mission is to transform the drug regulatory operations with the customer focus through analytics and informatics. They work towards the advancement for the CDERs public health mission. >> What are some of the objectives that you guys have? What are some things you guys have as your core top objectives of the CDER, the drug research group? >> The core objectives is we wanted to make sure that we are promoting a safe use of the marketed drugs. We want to make sure there's the availability of the drugs that are going to the patients are effective. And also the quality of the drugs that are being marketed are able to protect public health. >> What are some of the challenges that you guys have to take in managing the pharmaceutical safety, because I can only imagine certainly now that supply chains, tracing, monitoring, drug efficacy, safety, all these things are happening. What are some of the challenges in doing all this? >> In our office there are challenges in three different areas. One is the drug regulation challenges because as drugs are being more advanced and as there are more increasingly complex products, and there are challenging in the development area of the drugs, we wanted to make sure here we have a regulation that supports any advancement in science and technology. The other thing is also Congress is actually given new authorities and roles for the FDA to act. For example the Drug Quality and Security Act, which means any drug that's they want to track and trace all the drugs that goes to the public is they know who are the distributors, who are the manufacturers. Then you have the 21st Century Cures Act, and also the CARES Act package which was recently assigned, which also has a lot of the OTC drug regulatory modernization. Then there's also the area of globalization because just as disease don't have any borders, Product safety and quality are no longer on one country. It's basically a lot of the drugs that are being manufactured are overseas and as a result we wanted to make sure there are 300 US ports. And we want to make sure the FDA regulated shipments are coming through correctly to proper venues and everything is done correctly. Those are some the challenges we have to deal with. >> So much going on a lot of moving purchase as people say, there's always drug shortages, always demand, knowing that and tracking it. I can only imagine the world you're living in because you got to be innovative, got to be fast, got to be cutting edge, got to get the quality right. Data is super critical. And can you share take a minute to explain some of the data challenges you have to address and how you did that. Because I mean I could almost just my mind's blown just thinking about how you live it every day. Can you just share some of those challenges that you had to address and how did you do? >> Some of the key challenges we actually see is we have roughly 170,000 regulatory submissions per year. There are roughly 88,000 firm registration and product listing that comes to us, and then there are more than 2 million adverse event reports. So with all these data submissions and organization as such as us we need it, we have multiple systems where this data is acquired and each has its own criteria for validating the data. Adding to it are internal and external stakeholders also want certain rules and the way the data is being identified. So we wanted to make sure there is a robust MDM framework to make sure to cleanse and enrich and standardize the data. So that it basically make sure the trust and the availability and the consistent of the data, is being supplied to published to the CDER regulatory data users. >> You guys are dealing with- >> Otherwise like it's almost to give them a 360 degree view of the drug development lifecycle. Through each of the different phases, both pre market which is before the drug hits the market, and then after it hits the market. We still want to make sure the data we receive still supports a regulatory review and decision making process. >> Yeah, and you got to deliver a consumer product to get people at the right time. All these things have to happen, and you can see it clearly the impacts everyday life. I got to ask you that the database question 'cause the database geek inside of me is just going okay. I can only imagine the silos and the different systems and the codes, because data silos is big document. We've been reporting on this on theCUBE for a long time around, making data available automation. All these things have to happen if there's data availability. Can you just take one more minute talk about some of the challenges there because you got to break down the silos at the same time you really can't replace them. >> That's true. What we did was we did leave it more of us I mean, step back like seven years ago, when we did the data management. We had like a lot of silo systems as well. And we wanted to look at we wanted to establish a, we knew we wanted to establish a master data management. So we took a little bit more of a strategic vision. And so what we ended up saying is identifying what are the key areas of the domain that will give us some kind of a relationship. What are the key areas that will give us the 360 degree lifecycle? So that's what we did. We identified the domains. And then we took a step back and said and then we looked at what is the first domain we wanted to tackle. Because we know what are these domains are going to be. And then we were like, okay, let's take a step back and say which is the domain we do it first that will give us the most return on investment, which will make people actually look at it and say, hey, this makes sense. This data is good. So that's what we ended up looking at. We looked at it as at both ends. One is from a end user perspective. Which is the one they get the benefit out of and also from a data silo perspective which is the one data domains that are common, where there's duplication that we can consolidate. >> So that's good. You did the work up front. That's critical knowing what you want to do and get out of it. What were some of the benefits you guys got out of it. From an IT standpoint, how does that translate to the business benefits? And what was achieved? >> I think the benefits we got from the IT standpoint was a lot of the deduplication was not theirs. Which basically means like a lot of the legacy systems and all of the manual data quality work we had to do we automated it. We had bots, we also had other automation process that we actually put into work with Informatica, that actually helped us to make sure it's the cost of it actually went for us considerably. For example it used to take us three days to process submissions. Now it takes us less than 24 hours to do it, for the users to see the data. So it was a little bit more, we saw the, we wanted to look at what are the low hanging fruits where it's labor intensive and how can we improve it. That's how we acted there. >> What are some of the things that you're experiencing? I mean, like, we look back at what it was before, where it is now? Is it more agility, you more responsive to the changes? Was it an aspirin? Was it a complete transformation? Was some pain reduced? Can you share just some color commentary on kind of before the way it was before and then what you're experiencing now? >> So for us, I think before, we didn't know where the for us, I mean, I wouldn't say we didn't know it, when we have the data, we looked at product and it was just product. We looked at manufactured they were all in separate silos. But when we did the MDM domain, we were able to look at the relationship. And it was very interesting to see the relationship because we now are able to say is. for example, if there is a drug shortage during due to hurricane, with the data we have, we can narrow down and say, Hey, this area is going to be affected which means these are the manufacturing facilities in that area , that are going to be not be able to function or impacted by it. We can get to the place where the hurricane tracks we use the National Weather Service data, but it helps us to narrow down some of the challenges and we can able to predict where the next risk is going to be. >> And then before the old model, there was either a blind spot or you were ad hoc, probably right? Probably didn't have that with you. >> Yeah, before you were either blind or you're doing in a more of a reactionary not proactively. Now we are able to do a little bit more proactively. And even with I mean drug shortages and drug supply chain are the biggest benefit we saw with this model. Because, for us the drug supply chain means linking the pre and post market phases that lets us know if there's a trigger and the adverse events, we actually can go back to the pre market side and see where the traceability is who's at that truck. What are all the different things that was going on. >> This is one of the common threats I see in innovation where people look at the business model and data and look at it as a competitive advantage, in this case proactivity on using data to make decisions before things happen, less reactivity. So that increases time. I mean, that would probably you're saying, and you get there faster, if you can see it, understand it, and impact the workflows involved. This is a major part of the data innovation that's going on and you starting to see new kinds of data whereas has come out. So again, starting to see a real new changeover to scaling up this kind of concept almost foundationally. What's your thoughts just as someone who's a practitioner in the industry as you start to get this kind of feelings and seeing the benefits? What's next, what do you see happening because you haven't success. How do you scale it? What how do you guys look at that? >> I think our next is we have the domains and we actually have the practices that we work. We look at it as it's basically data always just changes. So we look at is like what are some of the ways that we can improve the data? How can we take it to the next level. Because now they talk about power. They are also warehouse data lakes. So we want to see is how can we take these domains and get that relationship or get that linkages when there is a bigger set of data that's available for us. What can we use that and it actually we think there are other use cases we wanted to explore and see what is the benefit that we can get a little bit more on the predictability to do like post market surveillance or like to look at like safety signals and other things to see what are the quick things that we can use for the business operations. >> It's really a lot more fun. You're in there using the data. You're seeing the benefits and real. This is what clouds all about the data clouds here. It's scaling. Super fun to talk about and excited. When you see the impacts in real time, not waiting for later. So congratulations. You guys have been selected and you receive recognition from Informatica as the 2020, Intelligent Disrupter of the year. congratulations. What does that mean for your organization? >> I think we were super excited about it. But one thing I can say is when we embarked on this work, like seven years ago, or so, problem was like we were trying to identify and develop new scientific methods to improve the quality of our drugs to get that 360 degree view of the drug development lifecycle. The program today enables FDA CDER to capture all the granular details of data we need for the regulatory data. It helps us to support the informed decisions that we have to make in real time sometimes or and also to make sure when there's an emergency, we are able to respond with a quick look at the data to say like, hey this is what we need to do. It also helps the teams. It recognizes all the hard work. And the hours we put into establishing the program and it helped to build the awareness within FDA and also with the industry of our political master data management is. >> It's a great reward to see the fruits of the labor and good decision making I'm sure it was a lot of hard work. For folks out there watching, who are also kind of grinding away in some cases, some cases moving faster. You guys are epitome of a supply chain that's super critical. And speed is critical. Quality is critical. A lot of days critical. A lot of businesses are starting to feel this as part of an integrated data strategy. And I'm a big proponent. I think you guys have have a great example of this. What advice would you have for other practitioners because you got data scientists, but yet data engineers now who are trying to architect and create scale, and programmability, and automation, and you got the scientists in the the front lines coming together and they all feed into applications. So it's kind of a new things go on. Your advice to folks out there, on how to do this, how to do it right, the learnings, share. >> I think the key thing I, at least for my learning experience was, it's not within one year you're going to accomplish it, It's kind of we have to be very patient. And it's a long road. If you make mistakes, you will have to go back and reassess. Even with us, with all the work we did, we almost went back a couple of the domains because we thought like, hey, there are additional use cases how this can be helpful. There are additional, for example, we went with the supply chain, but then now we go back and look at it and say like, hy, there may be other things that we can use with the supply chain not just with this data, can we expand it? How can we look at the study data or other information so that's what we try to do. It's not like you're done with MDM and that is it. Your domain is complete. It's almost like you look at it and it creates a web and you need to look at each domain and you want to come back to it and see how it is you have to go. But the starting point is you need to establish what are your key domains. That will actually drive your vision for the next four or five years. You can't just do bottom up, it's more of like a top down approach. >> That's great. That's great the insight. And again, it's never done. I mean, it's data is coming. It's not going away. It's going to be integrated. It's going to be shared. You got to scale it up. A lot of hard work. >> Yeah. >> Shanthi thank you so much for the insight. Congratulations on your receiving the Disrupter of the Year Award winner for Informatica. congratulations. Intelligence >> Yeah, thank you very much for having me. Thank you. >> Thank you for sharing, Shanthi Vigneshswaran is here, Office of Strategic programs at the Center for Drug Evaluation and Research with the US FDA. Thanks for joining us, I'm John Furrier for theCUBE. Thanks for watching. (soft music)

>> fly from San Francisco. It's the Cube covering Google Club next nineteen Rock Tio by Google Cloud and its ecosystem Partners. >> Okay, welcome back, everyone. We're here live in San Francisco for cubes coverage of Google next twenty nineteen. I'm suffering my coast, David. Want to many men also doing interviews out, getting, reporting and collecting all the data. And we're gonna bring it back on the Q R. Next to gas mark in l. A. Who's a senior technical account manager? AccuWeather at enough was the director product manager. Google Cloud Platform. Now welcome back to the Cube and >> thank you for >> coming on. Thank you. >> You got a customer. Big customer focus here this year. Step function of just logo's growth. New announcements. Technical. Really good stuff. Yeah. What's going on? Give us the update. AP economies here, full throttle. >> I mean, you know, the great thing is it's a pea eye's on all fronts. So what you saw this morning was about standardizing the AP eyes that cloud infrastructure is based on. You saw, You know, how do we build applications with AP eyes at a finer grained level? Micro services, you know, And we've had a lot of great customer examples of people using, and that's what you know with AC. You weather here talking about how do you use a P ice to service and build business models reached developer ecosystems. So you know. So I look at everything today. It's every aspect of it brings it back home tape. Yas. >> It's just things that's so exciting because we think about the service model of cloud and on premise. And now cloud, it's integration and AP Eyes or Ki ki and all and only getting more functional. Talk about your implementation. Aki weather. What do you guys do with Apogee? Google clouds just chair. What >> would implementation is so accurate? There's been running an AP I service for the past ten years, and we have lots of enterprise clients, but we started to realize we're missing a whole business opportunity. So we partnered with Apogee, and we created a new self survey P developer portal that allows developers to go in there, sign up on their own and get started. And it's been great for us as far as like basically unlocking new revenue opportunities with the FBI's because, as he said, everything is a p i cz. We also say everything is impacted by the weather. So why not have everyone used ac you other empty eyes to fulfill their weather needs? >> It wasn't like early on when you guys were making this call, was it more like experimenting? Did men even have a clue where they're like You's a p I I was gonna start grass Roots >> Way knew right >> away like we were working very heavily with the enterprise clients. But we wanted to really cater to the small business Is the individual developers to weather enthusiasts are students. Even so, we wanted to have this easy interface that instead of talking to a sales rep, you could just go through this portal and sign upon your own. It get started and we knew right away there is money to be left or money to be had money left on the table. So we knew right away with by working with apogee and creating this portal, it would run itself. Everyone uses a P eyes and everyone needs to weather, so to make it easier to find and use >> and what was it like? Now let's see how >> it we've been using it now for about two years, and it's been very successful. We've we've seen great, rather revenue growth. And more importantly, it's worked as a great sales channel for us because now, instead of just going directly to an enterprise agreement and talking about legal terms and contracts, you can go through this incremental steps of signed up on your own. Do a free trial. Then you could buy a package. You can potentially increase your package, and we can then monitor that. Let them do it on their own, and it allows us ability to reach out to them and see could just be a new partner that we want to work with, or is there a greater opportunity there? So it's been great for us as faras elite generator in the sales channel to really more revenue, more opportunities and just more aware these'LL process a whole new business model. It's amore awareness, actually replies. Instead, people were trying to find us. Now it's out there and people see great Now it Khun, use it, Get started >> Admission in the back end. The National Weather Service, obviously the government's putting up balloons taking data and presumably and input to your models. How are they connecting in to the AP eyes? Maybe described that whole process. Yeah. Tilak, You other works >> of multiple weather providers and government agencies from around the world. It's actually one of our strengths because we are a global company, and we have those agreements with all kinds of countries around the world. So we ingest all of that data into our back and database, and then we surface it through our story and users. >> Okay, so they're not directly sort of plugging into that ap economy yet? Not yet. So we have to be right there. Well, I >> mean, for now we have the direct data feeds that were ingesting that data, and we make it available through the AC you other service, and we kind of unjust that data with some of our own. Augur those to kind of create our own AccuWeather forecast to >> That's actually a barrier to entry for you guys. The fact that you've built those pipelines from the back end and then you expose it at the front end and that's your business model. So okay, >> tell about that. We're where it goes from here because this is a great example of how silly the old way papering legal contracts. Now you go. It was supposed to maybe eyes exposing the data. Where does it go from here? Because now you've got, as were close, get more complex. This is part of the whole announcement of the new rebranding. The new capabilities around Antos, which is around Hey, you know, you could move complex work clothes. Certainly the service piece. We saw great news around that because it gets more complex with sap. Ichi, go from here. How did these guys go? The next level. >> So, you know, I think that the interesting thing is you look at some of the themes here that we've talked about. It's been about unlocking innovation. It's about providing ways that developers in a self service way Khun, get at the data. The resource is that they need ask. They need them to build these types of new types of applications and vacuum weather experience and their journey on. That's a great example of it. Look, you know, moving from from a set of enterprise customers that they were serving very well to the fact that really ah, whole ecosystem of applications need act needs access to weather data, and they knew that if they could just unlock that, that would be an incredibly powerful things. So we see a lot of variants of that. And in fact, a lot of what you see it's on announcements this morning with Google Cloud is part of that. You know, Google Cloud is very much about taking these resource is that Google is built that were available to a select few and unlocking those in a self service fashion, but in a standard way that developers anywhere and now with andthe oh, switches hybrid a multi cloud wherever they are being able to unlock those capabilities. So why've you? This is a continuation of a P. I promise. You know, we're very excited about this because what we're seeing is more and more applications that are being built across using AP eyes and more more environments. The great thing for Apogee is that any time people are trying to consume AP eyes in a self service fashion agile way, we're able to add value. >> So Allison Wagner earlier was we asked her about the brand promise, and she said, We want our customers, customers they're not help them innovate all the way down our customers customers level. So I'm thinking about whether whether it gets a bad rap, right? I mean, >> look at it >> for years and we make make jokes about the weather. But the weather has been looked uncannily accurate. These they used to be art. Now it's becoming more silent. So in the spirit of innovation, talk about what's happening just in terms of predicting whether it's, you know, big events, hurricanes, tornadoes and some of the innovation that's occurring on that end. >> Well, I mean, look at from a broader standpoint to weather impacts everything. I mean, as we say, you look at all the different products out there in the marketplace that use whether to enhance that. So there's things you can do for actionable decisions, too. It's not just what is the weather, it is. How can whether impact what I'm doing next, what I'm doing, where I go, what I wear, how I feel even said every day you make a conscious and subconscious decision based on the weather. So when you can put that into products and tools and services that help make those actionable decisions for the users. That makes it a very, very powerful products. That's why a lot of people are always seeking out whether data to use it to enhance their product. >> Give us an example. >> What So a famous story I even told Justin my session earlier. Connected Inhaler Company named co hero they use are FBI's by calling our current conditions every time a user had a respiratory attack over time, it started to build a database as the user is using your inhaler. Then use machine learning to kind of find potential weather triggers and learn pattern recognition to find in the future. Based on our forecast, a p I When white might that user have another attack? So buy this. It's a connected health product that's helping them monitor their own health and keep them safe and keep them prepared as opposed to being reactive. >> The inhaler is instrumented. Yeah, and he stated that the cloud >> and that's just that's just one product. I mean, there's all kinds of things connected, thermostats and connect that >> talks about the creativity of the application developer. I think this highlights to me what Deva is all about and what cloud and FBI's all about because you're exposing your resource products. You don't have to have a deaf guy going. Hey, let's car get the pollen application, Martin. Well, what the hell does that mean? You put the creativity of the in the edge, data gets integrated to the application. This kind of kind of hits on the core cloud value problems, which is let the data drive the value from the APP developer. Without your data, that APP doesn't have the value right. And there's multiple instances of weird what it could mean the most viable in golf Africa and Lightning. Abbott could be whatever. Exactly. So this is kind of the the notion of cloud productivity. >> Well, it's a notion of club activity. It's also this idea of a digital value change. So, you know, Data's products and AP Icer products. And and so now we see the emergence of a P I product managers. You know, you know this idea that we're going to go and build a whole ecosystem of products and applications, that meat, the whole set of customer needs that you might not even initially or ever imagine. I'm sure you folks see all the time new applications, new use cases. The idea is, you know, can I I take this capability or can I take this set of data, package it up us an a p I that any developer can use in anyway that they want to innovate on DH, build new functionality around, and it's a very exciting time in makes developers way more productive than they could have been in >> this talks about the C I C pipeline and and programmable bramble AP eyes. But you said something interesting. I wanna unpack real quick talk about this rise of a pipe product managers because, yes, this is really I think, a statement that not only is the FBI's around for a long time to stay, but this is instrumental value. Yes. What is it? A byproduct. Men and okay, what they do. >> So it's a new concept that has Well, I should say a totally new concept. If you talk to companies that have provided a P eyes, you go back to the the early days of you know, folks like eBay or flicker. All of these idea was that you can completely reinvent your business in the way that you partner with other companies by using AP eyes to tie these businesses together. And what you've now seen has been really, I'd say, over the last five years become a mainstream thing. You've got thousands of people out there and enterprises and Internet companies and all sorts of industries that are a P I product managers who are going in looking at how doe I packet a package up, the capabilities the business processes, the data that my business has built and enable other companies, other developers, to go on, package these and embed them in the products and services that they're building. And, uh, that's the job of a P A. Product measures just like a product manager that you would have for any other product. But what they're thinking about is how do they make their A P? I success >> had to Mark's point there. He saw money being left on the table. Small little tweak now opens up a new product line at an economic model. The constructor that's it's pretty *** good. >> It's shifting to this idea platform business models, and it's a super exciting thing that we're seeing the companies that successfully do it, they see huge growth way. Think that every business is goingto have to transition into this AP I product model eventually. >> Mark, what's the what's the role of the data scientist? Obviously very important in your organization and the relationship between the data scientists and the developers. And it specifically What is Google doing, Tio? Help them coordinate, Collaborate better instead of wrangling data all day. Yeah, I mean, >> so far, a data scientists when we actually have multiple areas. Obviously, we're studying the weather data itself. But then we're studying the use case of data how they're actually ingesting it itself, but incorporating that into our products and services. I mean I mean, that's kind of >> mean date is every where the key is the applications have the data built in. This is to your point about >> unnecessarily incorporating it in, but to collaborate on creating products, right? I mean, you're doing a lot of data science. You got application developers. All right? You're talking about tooling, right? R, are they just sort of separate silos or they >> I mean, we obviously have to have an understanding of what day it is going to be successful. What's gonna be adjusted and the easiest way to adjust it a swell so way obviously are analyzing it from that sense is, >> I say step back for a second. Thiss Google Next mark. What's your impression of the show this year? What's the vibe? What's this day? One storyline in your mind. Yet a session you were in earlier. What's been some of the feedback? What's what's it like >> for me personally? It's that AP eyes, power, everything. So that's obviously what we've been very focused on, and that's what the messaging I've been hearing. But yeah, I mean, divide has been incredible here. Obviously be around so many different great minds and the creativity. It's it's definitely >> talk. What was the session that you did? What was the talk about? Outside? Maybe I was some >> of the feedback. Yeah, I mean, so the session I gave was how wacky weather unlock new business opportunities with the FBI's on way. Got great feedback was a full house, had lots of questions afterwards that followed me out to the hallway. It's was actually running here, being held up, but lots people are interested in learning about this. How can they unlock their own opportunity? How can they take what they have existing on and bring it to a new audience? For >> some of the questions that that was kind of the thematic kind of weaken stack rank, the categorical questions were mean point. The >> biggest thing was like trying to make decisions about how for us, for example, having an enterprise model already transitioning that toe a self serve model that actually worked before we're kind of engaging clients directly. So having something that users could look at and on their own, immediately engage with and connect with and find ways that they can utilize it for their own business models and purposes. >> And you gotta be psychic, FBI as a business model, You got FBI product managers, you got you got the cloud and those spanning now multiple domain spaces on Prem Hybrid Multi. >> Well, that last points are very exciting to us. So, you know, if you look at it, you know, it was about two and a half years ago that apogee became part of Google and G C P got into hybrid of multi cloud with aptitude that we were, you know, the definitive AP I infrastructure for AP eyes. Wherever they live. And what we saw this morning was DCP doubling down in a very big way on hybrid of multi clap. And so this is fantastic four. This message of AP eyes everywhere. Apogee is going to be able Teo sit on top of Antos and really, wherever people are looking at either producing are consuming AP eyes. We'LL be able to sit on top of that and make it a lot easier to do. Capture that data and build new business models. On top of it, >> we'LL make a prediction here in the Cube. That happens. He's going to be the center. The value proposition. As those abs get built, people go to the business model. Connecting them under the covers is going to be a very interesting opportunity with you guys. It's >> a very exciting, very exciting for us to >> get hurt here first in the queue, of course. The cubes looking for product manager a p I to handle our cube database. So if you're interested, we're always looking for a product manager. FBI economies here I'm Jeopardy Volante here The Cube Day one coverage Google Next stay with us for more of this short break

>> Narrator: From Las Vegas, it's the Cube, covering IBM Think 2018, brought to you by IBM. >> Welcome back to Las Vegas, everybody. You're watching the Cube, the leader in live tech coverage. My name is Dave Vellante, and this is day three of our wall-to-wall coverage of IBM's inaugural Think conference. Mary Glackin's here, she's the vice president of weather business solutions, public, private partnerships, IBM Watson, and she's joined by Sheri Bachstein as the global head of consumer business at the Weather Company, an IBM company. Ladies, welcome to the Cube, thanks so much for coming on. >> Thank you, you're welcome. >> Thanks. >> Alright, Mary, going to start with the Weather Company. When IBM acquired the Weather Company, a lot of people were like, "What?", and they said, "Okay, data science, I get that.", and then, there was an IoT spin on that. Obviously, you have a lot of data, but, I got to ask you, what business are you in? >> So, what we like to say is we're in, not in the weather business, we're in the decision business. We're really dedicated, everyday, to help businesses, make the best decisions possible, and Sheri works on the consumer end of the business to do exactly the same thing. >> So, talk about your respective roles. Sheri, you're on the consumer side, as Mary just said, what does that entail? >> So, the consumer side is any touchpoint where we're bringing weather and weather insights to our consumers, whether it's on our weather channel app, whether it's on our web platform, mobile web, on wearables, so, it's anywhere where we're connecting with consumers, and, as Mary said, it's really about helping consumers make decisions. In our field, the forecast and some of the weather data has become a commodity almost, and we've actually shared our weather data with a lot of partners, and, so, now, we're using machine learning and data science to really come up with weather insights to help consumers make decisions, and it could be something just as simple as what to wear today, what's going to happen for a big event, or it can be around how do I keep people safe during severe weather. >> Yeah, I mean, we all look at the weather. I mean, I look at it everyday. >> Yeah. >> Of course, when you travel, like, what do I bring, what do I wear? Living in the East Coast these days, a lot of storms that we've >> That's right. >> encountered in the East Coast. I wonder if you could talk about life at IBM. I mean, again, it was a curious acquisition to a lot of people. Have you guys assimilated, how has it changed your business? >> I would say pretty dramatically. So, coming back to IBM acquiring us, they acquired us, really, for two reasons. One is we had some underlying technology that was really of interest to them that they're leveraging today, but the other part was because weather impacts so many businesses. So, as we've come into IBM, we've had alliances with IBM research. We're working on a pretty exciting project in bringing the next generation weather model to market, using high performance computing there. We've had alliances, definitely, through Watson in bringing AI into our products, and then, our product lines marry up with a lot of IBM product lines. So, we've rolled out a really exciting offering in closed captioning, and it really works well with some of the classical media business, weather media business that we have been providing. >> So, how do you guys make money? Maybe we could talk about the consumer side and the business side. A lot of people must ask that question. >> Yeah. >> They're advertising, okay, fine, >> Yeah. >> but that's not the core of what you guys do. >> Yeah, so, on the consumer side, a big majority of our revenue is drive by advertising, but we had to look at that business as well, 'cause as programmatic advertising has kind of taken up the landscape, how did we pivot to really generate more revenue, and, so, we've done that by creating Watson advertising, and that was one of the first implementations of Watson after the acquisition on the consumer side, and what we've done is we've created an open, scalable environment that, now, we can not only sell meaningful insights on our platform, but we can now give that to our partners, that they can go off our property and use the weather insights, we can use different data around location and media to help our partners really have a better experience, not only on our platform, but on any publisher's platform. >> So, that's your customers using Watson for advertising to drive their business. >> That's right. >> It's not like IBM is getting into the advertising business, per se, directly, is that right? >> Right, well, we're leveraging the power of Watson to create these insights. One of the products we created is called Weather FX, and, really, what it's doing, it's taking predictive analytics on the retail side, which is really an underused technology for retailers, but taking our historical weather data, mixing it with their retail data' to come up with insights so we can come up with interesting things that, say, in the northeast, like right now, during the winter, soda sells tremendously during very snowy or rainy winters. We can look at, you know, strawberry Pop-Tarts sell fairly well right before a hurricane, and, so, these are insights that we can bring to retailers, but it helps them with their supply chain, it helps them with their inventory, it can actually even help them with pricing, and, so, this is one of the ways we're taking our weather technology and marrying it with the advertising world to help provide those insights. >> For real, with the strawberry Pop-Tarts? >> For real, yeah, I guess, you know, you don't have to cook 'em or something. I don't know, so, yeah. >> Right, yeah, it's simple if the lights go out, okay. I mean, we want to ask you about your title, public and private partnerships. It's interesting, what is that all about? >> So, it's really about the fact that weather has really been something that's been shared globally around the world for hundreds of years at this point, and, so, the Weather Company and IBM take it very seriously that we be good partners in that community of weather providers. So, one of the things that we feel passionately about is we have a shared safety mission with national meteorological services globally. So, here in the US, we transmit, Sheri's team does, the warnings that come from the National Weather Service unaltered with attribution to the National Weather Service. We feel that it's really important that there's a sole authoritative voice when there's really danger. So, we share that safety mission, and then, we're trying to help in other parts of the world. We've had some partnerships to try to increase the observing in Africa which is really a part of the world that's under-observed. So, some of IBM's philanthropic efforts have been helping to fill in there and work with those national met services. So, it's really one of the really fun parts of my job. >> You know, we talk a lot about digital transformation, and Ginni Rometty was talking about the incumbent disruptors, and we've been riffing on that all week. We've made the observation that companies that are digital have data at their core, and they've organized, sort of, human expertise around that data. Most companies, Fortune 1000, are built around human expertise and built around other assets, the bottling plant or the factory, et cetera. I look at the Weather Company as a data company, that's probably fair. Did you evolve into that data is clearly at your core? Has it always been, and it's very interesting that IBM has acquired this company as it changes its DNA. I wonder if you could address that. >> Go ahead (laughs). >> So, I think there's a couple aspects around our data. There's obviously the weather data which is really powerful, but then, there's also location data. We're one of the largest location data providers besides Google and some of the others, because our weather accuracy starts with location which is really important. We have 250 million users that use our application, and we want to give them the most accurate forecast, and that starts with location. Because we add value, users will opt in to give us that data which is really important to us that we do keep their data private and opt in to that to get that location data. So, that's really powerful, because, now we can deliver products based on time and location and weather, and it just makes for better weather insights for, not only our consumers, but for our businesses. >> Yeah, yeah. >> Do you use, I mean, how do you use social? I mean, you know how Waze tells you where the traffic is and you report back. Do you guys rely heavily on that, or do you more rely on machines to help you with your forecast? Is it a combination? >> So, I could talk a little bit. One of our new market areas we've been going into is ground transportation. So, we do have a partner that's providing us some transportation, traffic information, but what we bring to it is being able to do, the predictive thing, is to take the weather piece and how that's going to influence that traffic. So, as the storm comes through, we know by looking at past events what that will mean and we bring that piece to the table. So, it's an example of how we go, not just giving you a weather forecast, but really forecasting the impacts and giving you insights, so that if you're running a large trucking operation, you can reroute fleets around it and avoid weather like that and keep people safe. >> Talk about, oh, go ahead, please. >> One of the brands within our portfolio is Weather Underground, and what they brought to the table for us is a personal weather station that works. So, we have about 270,000 around the world, and these are people that just really love the weather. They have a personal weather station in their backyard and they provide that data that then goes into Mary's team in helping looking at the forecast. So, that's one of the ways that we're using kind of a social network in sensoring to influence some of the work that we're doing. >> I mean, the weather forecast, for years, have been the butt of many jokes. You guys are data science oriented, data scientists, the data doesn't lie. We just keep iterating >> Yeah. >> and make it better and better and better. What could you tell us about the improvements of the forecast over the last decade? Maybe Bill Belichick makes jokes about the weather and you hear it, you say, "You know, actually "the weather's predictions have gotten much better." You guys measure it, what can you share with us? >> Oh, it's gotten so much better over the course of my career, it's pretty dramatic and it's getting better still. You're going to see some real breakthroughs coming up. So, one of the things that we've really put a lot of bets on in IBM is the internet of things, >> Dave: Right. >> and, so, we are, today, pulling off of cellphones atmospheric pressure data and that's going into our next generation model. So, this'll be more data than anybody has powering that model. So, you're able to augment traditional data sources like, you may or may not know, we still launch weather balloons twice a day to measure through the atmosphere, but, in our technology, we take data off of airplanes, we take data off of cellphones, we'll soon be taking data off of cars which will tell us when the windshield wipers are moving, is it raining or not, when the anti-lock brakes things lock, that roads are icy, all of that. So, all of that will come in to improve forecasting. >> So, this requires partnerships with all that and amazing supply chain. >> Absolutely. >> I presume IBM helps there as well, but did you have a lot of that in motion prior to the acquisition, how does that all work? >> I think we've really been empowered by IBM. >> Yep, absolutely. >> Yeah. >> There's no question about that, and it's about finding the win-win. When we work with car manufacturers they're looking to have safe experiences for their drivers and we can help in that regard, and, as we move into autonomous vehicles, there's just going to be even more demand for very high resolution, accurate weather information. >> Am I correct at all, the weather data from all these devices actually goes back to the IBM cloud, is that right, and that's where the models are iterated and developed, is that correct, or does some of it stay out in the network? >> It's all a cloud-based operation that's here. We do do some, I mentioned before that we're working with IBM research on next generation high-performance computing which is actually, it can be cloud-based, but it's also on Prim-based, because of the very large cores we need for computing these models. We're going to run a very high-resolution model globally at a very high frequency. >> So, thinking about some of the industries that you're helping, I mean, you mentioned retail before. Obviously, government's very interested in this. I would imagine investors are interested in the weather in a big way. >> Yeah. >> Maybe you could talk about some of the more interesting industries, use cases, business models. >> Yeah, there's a lot out there, there's traditional ones we've served for years like energy traders that are very interested in, you know, because they're trying to make decisions about that. The financial services sector is also very interested. When they can get some additional insights through footfall traffic, if they know certain stores are seeing more footfall traffic, that will give them some indication, a little edge up in the marketplace for that. So, we see those kind of things, and other traditional areas as well, agriculture, what you would expect there. >> So people, you know, you hear a lot of talk in the press about artificial intelligence and Elon Musk predictions and the like, but here's an example where machine intelligence, everybody welcomes, keeps getting better and better and better. How far could we take AI and weather? Where do you see this going in the next 10 years? >> So, on the consumer side, I think it's really about transforming the way that we're delivering weather on the digital platform, the new age of the weather app will say, and, really, users want a personalized experience. They want to know how the weather's going to impact me, but they don't want to personalize, right? So, that's where machine learning is coming in, that we can be able to provide those insights. We'll know that, maybe, you're an allergy sufferer or migraine sufferer, and we're going to tell you that the conditions are right for that you might have symptoms related to that around health. So, there's a lot of ways, on the consumer side, more personalized experience, giving you more assurance that you don't have to, necessarily, go to the app to find information. We're going to send it to you more proactively, and, so, machine learning is helping us do that cognitive science as well. So, it's a pretty exciting time to be part of the weather. >> Yeah, that bum knee I have, you know, you might want to get ahead of the pain. >> That's right, with the arthritis, yes, yes, so, definitely. >> Alright, Mary, we'll give you last word on IBM Think and, you know, the whole trend of AI and weather. >> So, I think it's really exciting. I think Ginni says it really well. It's about AI and the person as well. You know, AI doesn't take over. It's really finding the way to AI to really assist decision makers and that's we're going on the business end of things is really sorting through tons and tons of data to really provide the insights that people can make, businesses can make really great decisions. >> Well, it's always been a really fascinating acquisition to me, and, now, just to see how it's evolving is really amazing. So, Sheri and Mary, thanks very much for coming on the Cube >> Thank you. >> and sharing your experiences. >> Thanks so much. >> Great, thank you. >> You're welcome, alright, keep it right there, everybody, you're watching the Cube. We're live from Think 2018 and we'll be right back. (techno beat)