Microsoft Research Podcast - 031 (rerun) - AI for the Developing World with Dr. Ranveer Chandra

Episode Date: July 4, 2018

This episode first aired in April (2018). When we think about artificial intelligence and the “world of the future,” our vision is usually more Jetsons than Green Acres. But for Dr. Ranveer Chandra, a Principal Researcher in the Systems and Networking group at Microsoft Research, rural farms are the perfect place to realize the benefits of AI through what he calls precision agriculture, or data-driven farming.Today, in a wide-ranging interview, Dr. Chandra talks about how his research may eventually make your wi-fi signal stronger and your battery life longer, but also shares the story of how spending childhood summers with his grandparents in rural India inspired a line of research that could change the face of farming and help meet the food and nutrition needs of a growing global population.

Hey there, Microsoft Research Podcast listeners. For the summer break, we're bringing back a curated selection of previously aired podcasts. Whether you're downloading for a road trip or just tuning in for the Wednesday commute, we invite you to listen again or discover a new researcher for the first time. Enjoy. One of the latest projects I'm doing is Farm Beats.

And for that, to actually see how well your research works, you need to be out there in a farm. And a lot of times when I'm going there, it's raining out here in Seattle. I'm like, what am I doing going out there in the farm in the middle of nowhere in this rain? But on the other hand, just reminding yourself that if this research works, this is going to benefit so many farmers, gives you that level of energy to keep going, to keep thinking about doing bigger things and not stopping where you are. You're listening to the Microsoft Research Podcast, a show that brings you closer to the cutting edge of technology research and the scientists behind it. I'm your host, Gretchen Huizinga.

When we think about artificial intelligence and the world of the future, our vision is usually more Jetsons than Green Acres. But for Dr. Ranveer Chandra, a principal researcher in the Systems and Networking Group at Microsoft Research, rural farms are the perfect place to realize the benefits of AI through what he calls precision agriculture or data-driven farming. Today, in a wide-ranging interview, Dr. Chandra talks about how his research may eventually make your Wi-Fi signal stronger and your battery life longer, but also shares the story of how

spending childhood summers with his grandparents in rural India inspired a line of research that could change the face of farming and help meet the food and nutrition needs of a growing global population. That and much more on this episode of the Microsoft Research Podcast. Ranveer Chandra, welcome to the podcast. It's great to see you today. Yeah, thank you. I'm excited to be here. So you're a principal researcher in systems and networking, and it says on the website that you develop technologies that expand the state of the art in computing. That's a pretty broad statement. What does it mean? I come from a networking background. I do computer networking, innovations that can help you connect devices better. That's what my PhD was in. My PhD was on a technology

called virtual Wi-Fi. Since 2009, 2010, I've started expanding the scope of what I do from networking to other technologies as well. For example, batteries and improving the battery life of devices and more recently to farms. In some sense, they are all about systems. They are all about how do you make technology better, more accessible? How can you take a big step in the different systems components, be it networking, be it operating systems, be it batteries, be it the Internet of Things? How do you connect devices to the Internet? And coming up with technologies that take the state of the art to the next level. And with each one of them,

that's what we are trying to accomplish. Yeah. And having experienced technology for decades now, I understand that connectivity, networking, making everything work together is one of the biggest challenges because you've got people all over the world coming up with new ideas, and how do you connect them all? Yeah. Yeah. And just to put that point in perspective, around half the world is still not connected to the internet, right? And that's what drives this point even more that we need innovations. We need to come up with new technologies to get everyone connected, to bring the benefits of technology to everyone, right? To democratize technology. Yeah. Ranveer, you're a busy guy. You have multiple research projects on the go. You've published more than 80 academic papers

that have been cited multiple times, and you've got more than 85 patents. And you're active in the larger research community. You're serving on committees. You've got projects, initiatives, awards, media, mentoring, advising PhDs. I mean, seriously, the list could go on. What do you do in your spare time?

Well, whatever time I get, I like thinking about the research problems, the open problems. That's what gives me the energy to keep going on everything else. I do like doing all the work you just mentioned, like being involved with students. That gives me a lot of excitement. That makes me happy. It just makes me think more about the research that I do. It gives me new ideas working with the younger community. That's why I serve on PhD committees.

I love talking to students. Whenever I visit a university, I ask my host to set up at least one hour talking to graduate students. So this is something that excites me. Same with in the research community. I like serving on program committees because it keeps me up to date with what's going on and thinking about new ideas. But in the spare time, yeah, I like thinking about these research problems. I do have a wife and a daughter and I like my family. We like doing things together as well, usually on the weekends.

But on weekdays, yeah, it's a busy schedule. It's work. Yeah, it's work and also working with different parts of the world. We work with India, with China, so it takes up nearly all our time during the week, but it's exciting. I'm not complaining. It's something that I like doing. Well, what's funny was that I was going to ask the follow-up question, seriously, what inspires you and how do you sustain this level of work? And you've just answered that with what do you do in your spare time? Because that is what you do for joy, sounds like. Yeah, in the sense, I really like my work.

I'm really excited to come to work every day. Even I travel a lot, but even when I'm traveling, I don't complain. I really like doing research and thinking about these meeting students. Like, I'll give you an example. One of the latest projects I'm doing is Farm Beats. And for that, to actually see how well your research works, you need to be out there in a farm, try your research. So we work with Sean Stratman, who's a smallholder farmer in Carnation, which is 25 miles east of here. And a lot of times when I'm going there, it's raining out here in Seattle. I'm like, what am I'm going there, it's raining out here in Seattle.

I'm like, what am I doing going out there in the farm in the middle of nowhere in this rain? But on the other hand, just reminding yourself that if this research works, this is going to benefit Sean, benefit so many farmers, gives you that level of energy to keep going, to keep thinking about doing bigger things and not stopping where you are. Well, let's talk about some of the specific projects you're working on. We just started sort of there, but one that really interested me was this TV white space dynamic spectrum, spectrum opportunity program that you're doing called White Space Networking Project. So what kind of things can you do with spare spectrum, so to speak, and why is it important?

So to give you an idea of what the TV white spaces can enable is, imagine if you go to Best Buy, you buy a Wi-Fi router and you plug it into your house. Imagine if a future router of this kind, you could access the Internet a few miles away. Right now, as soon as you exit your house, your Wi-Fi disappears. So the way we do that, this was work we did back in 2009. This was the best paper at ACM sitcom. The particular technology that we had developed here was a way to take a Wi-Fi signal and put it in an empty TV channel. What we mean by this is this is over the air TV, the TV you watch using antennas. So, you know, when you browse over the air TV, the TV you watch using antennas. So, you know, when you browse over the air TV, you see certain channels, you know, channel seven, you're getting a TV broadcast. On channel eight, all you see is white noise, nothing coming there.

The particular technology we developed was a way to put a Wi-Fi signal in that noise channel such that it doesn't interfere with your TV reception on an adjacent channel. So you could be watching channel seven. On channel eight, we could be sending Wi-Fi. And the reason this is so cool is that compared to Wi-Fi at the same power level, typically in UHF TV frequencies your signals will go four times farther. In VHF they'll go 12 times farther. And this is in free space, just really based on physics. If you put trees, crops, canopies, buildings, your signals propagate even more. In fact, in some of our experiments in India, where we connected a rural high school to the internet, our non-line of sight link was 10 miles. That's 16 kilometers.

We were able to stretch the internet by 16 kilometers to that high school, which previously didn't have any internet connection before. So that's what the TV white spaces enables. It enables you to send Wi-Fi signals over much longer distances. And there's a lot of technology behind that. If you go to our website, you'll find lots and lots of papers that we've written on it. But the FCC chairman was here in 2010 in this building, and we showed him this demo at that time.

And since then, this technology has been made legal in the U.S. And right now it's legal in several other countries as well. Can you just grab it? I mean, you don't have to go through any channels to apply for it or anything? No, so this is where a lot of the technology pieces were developed. There's the concept that at that time was proposed was called dynamic spectrum access. And now there's an alliance. And we have proposed an architecture where you have a web service in the background.

Your device would say, this is where I am. The web service would reply and say, at this location, these are the channels you can use. So that's a new way to establish an internet connection. But with that, we had shown that you can establish these connections in a way that doesn't interfere with the TV reception. So right now there is a big momentum around this. There are lots of companies involved in the Dynamic Spectrum Access Alliance. I'll be giving a keynote at their conference in London in May this year. And there is an entire ecosystem around the TV white spaces. New devices are getting built. The vision here is to be able to get the internet access to the remotest places in the world. And some of the messages you get from people like students in Africa sending a small message saying, thank you so much for this work. We were able to access the internet for the first time. These are really touching messages, which we think we're still not there with this research. We've made the first steps of connecting people, but that's the vision. We want to use this technology to get internet in the hands of everyone. You know, there's a text acronym SMH. Have you heard of it? No. It's called shaking my

head. And every single podcast, I'm shaking my head going, I cannot believe this is so interesting, so exciting and so useful. Right. Yeah. I mean, so at some point, can I just get extra white space if I want Wi-Fi while I'm driving? Because I'm always mad when I leave my house. At some point, will it be anyone? Yeah. So one of the things that happened since we did this work in 2009, the paper that we had written was called Wi-Fi, for Wi-Fi over the white spaces. So we would expect that in the future you would have these, like right now on your phone, you have 802.11abg, you could have, there's an IEEE 802.11 standard, 802.11af, which is also known as the Wi-Fi standard. And that is about taking the Wi-Fi standard and using it in the TV white spaces. So in the future,

who knows, this could be something that could be very real. Another area I was surprised to learn about was battery research, and you've referred to it a couple of times. Not something I would have expected be going on in Microsoft research for some reason. But tell us what's going on in battery research from your perspective and how it's going to make our lives better. I already think of 10 reasons. Yeah. So we started this research back in 2010, where I'm leading this battery research across Microsoft, where it's not just about research. We work with the product teams as well. And we are looking at ways in which you can improve the battery life of mobile devices, take it to the next level. Our mission is to keep working on this until you can get the battery life of mobile devices, take it to the next level.

Our mission is to keep working on this until you can get your devices to last a week without charge. And we're thinking of mobile devices. Mobile devices could be phones, laptops, tablets, two-in-ones, any of these. And we're taking a holistic approach, working all across the battery stack, from batteries itself, to how these batteries are managed in the hardware to the device drivers to the operating system to the applications all throughout and coming up with ways in which we can take it to

the next level to improve battery life and one of the things that we did initially was to come up with a way to model the battery consumption of the devices how much power does your display use, does your processor use, does your network use? We built models for all of them. And then we worked very closely with the different product teams on transferring the technology. And they're part of the research that we do, all the teams within Microsoft. One of the key technologies is software-defined batteries. And we're super excited about that. That's a new paradigm of the way you use batteries.

What we came up with was a way in which you can combine batteries of different chemistries. You can imagine a fast charging battery and a high energy density battery plugged together. And with a little bit of hardware, what we did was you could switch across the different batteries. And this is kind of the entire stack that we built all the way from batteries, the battery hardware, and the operating system module and the AI piece, the artificial intelligence piece of predicting what kind of workload is going to come. And based on that, we came up with this paradigm of software-defined batteries, based on which we showed that you can mix and match different batteries. So we did the first step with that of showing it on mobile devices, but the real passion

behind this comes from being able to use it in the grid. If you think of the electric grid or even electric vehicles, one of the biggest challenges there is how long is it going to last? How much farther can you go on one charge? Same with the electric grid. How much charge can you store? And until now, everyone still uses batteries of the same chemistry. We believe with this paradigm, which we developed a while back, the first time this idea came to us was in 2011. And since then, we kept modifying the technology before we published it in 2015. This was chosen as one of the research highlights in CACM, the communications of the ACM in 2016. And now we are thinking of broadening the scope of what we had developed to see where else can it be used. And you can imagine others in a, like I mentioned, electric vehicles, right? So whenever you sit in a car, if you enter your GPS coordinates, or even if you could learn that, you know what, now I'm going home, you know what kind of traffic it's going to hit, what kind of temperature it's going to be,

whether you're going to be braking a lot. All these things affect your battery consumption. So then based on where you're going, the prediction part is easier. You can easily build the smarts in so that your car can go that extra bit longer. Right, by pulling from a different source. By pulling from different batteries based on your profile,

based on what the traffic is, based on what the weather is and so on what the weather is, and so on. Okay, this is one of those podcasts where it's like everything you say, I can see how it can impact my life. That's my goal. Let's go a little bit more. I could talk to you about batteries for an hour. Seriously. And that's kind of weird for me.

I don't usually talk about batteries to people. One of the most well-known projects you're leading is an incubation on IoT applications called FarmBeats. You referred to it earlier, but it's really cool. I want to know more about it. I know our listeners would like to know more about it. So tell us about FarmBeats. Yes. Yeah, I'm super excited about it. I've been doing this for about three years now. So the reason I started this project is I don't have an agricultural background. You don't? I'm not a farmer. Could be one of those things on your resume.

I know. So my PhD is in computer science. But I did have some experience with farming in the sense that when I was growing up in India, as it happens in India, every summer and winter vacation, you have to spend with your grandparents. And my grandparents lived in a farm in a small village in Bihar. It's in rural India. And every summer and winter vacations, that's three months of summer vacation and one month of winter vacation, I would be spending there with them.

And I used to hate that. Because as it was at that time in the farms in India, the villages, they didn't have any electricity, no bathrooms. It wasn't easy living. And the rest of the time I was in a city. So I didn't like it, but I still went there. I spent a lot of time there. A lot of my friends were in the villages. I learned biking when I was there with those friends. Because they had to go too. Yeah, yeah, yeah. And so exactly. So, but then I think in a way that I didn't like farming then, but in a way it's kind of giving back. Having seen the poverty, the kind of the way people go about doing farming, I want to develop

technologies that can impact their lives. The goal of FarmBeats is to enable data-driven farming. So just to give you a bit of perspective, the world's food production needs to increase by about 70% by 2050 to feed the growing population of the world. And I was at another meeting with different people and there was a lady from the Rockefeller Foundation and she mentioned that, you know what, we're talking of 70% to feed the world, but what about nourishing the world? The challenge is even more giving them nutritious food. Good food. Yeah, good food. So how do we get there?

So it's like the green revolution all over again, because we need to meet this food production demand without an increase in the amount of arable land. The amount of arable land is about the same. The water levels are receding. The challenge is even more. We need to get there in a sustainable way. So agricultural scientists have been thinking about this problem. And the most promising technique seems to be that of data-driven farming. What we mean by data-driven farming is the ability to map a farm and overlay it with data. For example, what is my soil moisture level six inches below the soil? What is my soil nutrient level throughout the farm? If you can build maps like those, it can enable techniques such as precision agriculture. What we mean by precision agriculture is the ability to do site-specific applications. For example, rather than putting water uniformly throughout the farm, you apply it only where it is needed.

Instead of spraying pesticide throughout the farm, you apply it only where there are pests. Precision agriculture has been shown to improve yield, reduce cost. It's also better for the environment. And once you can build maps like this, the other benefit that we got approached by some other companies, which are like, we could use this to improve seed phenotypes. It's not only what precision agriculture could also help with phenotyping, which can help feed the world. Precision agriculture was proposed back in the 80s, that technique. It still hasn't taken off 30 years down the line. And one of the main reasons it hasn't taken off is because of the cost of data-driven agriculture solutions. Just to give you an idea, there was an expo that I had attended where there were lots of companies presenting their precision agriculture equipment and the cheapest sensors at that expo were five sensors for eight thousand dollars and a recurring cost for a farmer who doesn't make that much money most of them to invest in that kind of

infrastructure without knowing what the roi the return on investment is is asking for too much yeah so the goal of the farmbeats project is to bring down the cost of these data-driven agriculture solutions by two orders of magnitude. We want to bring it down from $8,000 to $80. And this is the goal. And we have some innovations based on which we think we can get there. Is there any connection to the White Space project and the FarmBeats project? Yeah. So one of the key reasons these existing sensors are expensive is because of the cost of connecting these sensors to the cloud. Imagine if you're out in the middle of the farm,

a lot of times you'll end up using satellite connectivity to connect them or custom cellular because a lot of times you don't get cellular connectivity, you'll build your own base stations. So the key insight in the context of agriculture was that TV towers are where people are in cities.

In Seattle, you have a lot of TV towers across the lake TV towers are where people are in cities. In Seattle, you have a lot of TV towers across the lake. The farms are away from the cities. There are fewer people in the farms. Because there are fewer people in the farms, you have fewer TV towers. What that means is you have a lot of empty TV channels. What that means is you have a lot of spectrum. We are talking of hundreds of megabits of unused capacity, at which point we're not only connecting sensors, we could be connecting drones, cameras, streaming a lot of data from the middle of the farm that you previously couldn't gather. Our vision is just like Wi-Fi connects your house, the TV

white spaces can be used to connect the farm. And in our deployments, that's what we do. We set up this antenna in the farm and miles around it now gets connected. You could be streaming data from the middle of nowhere. If you talk to agricultural scientists, the number one problem they face is that of getting data. So that's one of the innovations. The second thing we did was, so right now, if you wanted to build an accurate map, what is the soil moisture level,

six inches below the soil throughout my farm? Well, because soil moisture varies a lot, you would say, you know what, I'll deploy lots and lots of sensors. But deploying a lot of sensors is expensive to deploy, to manage. It also comes in the way of the farmer as the farmer does a day-to-day job. So then the key innovation we came up with was a way to build this map using very few sensors. And the way we do that is we combined ground sensor data with aerial imagery, for example, using these UAVs, the drones. So you fly a drone, get the visual image.

We take that to train a machine learning model using the ground sensor data. And then we use that with very few sensors. We're able to build very accurate maps of the farm. So using the TV white spaces, it brings down the cost of the sensors. Using this technology of fusing aerial imagery with drones, we need much fewer sensors than what you needed earlier. In the startup space, or even in research, you see a lot of people working on drones, a lot of people working on sensors. We believe we are the first who's been able to combine both of them in a meaningful way. We did that for soil moisture, soil temperature, and soil pH. Once we presented it to other companies, they want to use it for many other use cases. They

think this technology can now be used for many other applications in farming. The other innovation we came up with was with the white spaces, we can gather a lot of data. You can send a lot of data from the drones, from the cameras. You can be putting these cameras in barns, monitoring cows. You can send all that data to the farmer's house, but the farmer's house doesn't have great connectivity to the internet. So what we did was, our key insight was that most farmers have PCs. If they don't have a PC, we ship them an Azure IoT Edge device that sits on the farm. And this does a lot of compute, a lot of machine learning sitting on the farm itself. It can be available offline and so on,

and only sends the summaries to the cloud. And the fourth innovation we had was around power, where in Seattle, if I'm using solar panels, you need to come up with an innovation around power, especially in winter when you don't get a lot of sun. How do you get your system running? So we had some innovations there as well. But this is the complete package of things

that we call FarmBeats. We're able to gather a lot of data using the white spaces and then apply new machine learning on top of it using the technologies I mentioned to come up with insights. Yeah, this is going to be called the SMH podcast because I even did background research, but on you, this is fascinating and fantastic. Do you have to get the cow's permission to monitor them? That's a good question. We at least get the farmer's permission. We're going to watch your cows. I know. Well, it's for their own safety. I know and health and we can tell when a cow is sick and things like that. It's like a nursery monitor. Is the kid crying? Is the kid okay?

Let's talk a little bit about, I'm having so much fun here and all I'm doing is listening. Let's talk about tech transfer, which is kind of a broad term in the industry. It's probably the goal of most researchers to see what you described earlier, your passion to see your work have an impact on other people's lives. But you have a lot of stuff going on that's tech transfer. Is that part of your DNA with your work here in systems and networking? Yeah, and that was one of the reasons I had joined Microsoft Research in the first place. And when I graduated from Cornell in 2005, all other places I had interviewed in were academic positions. This was the only industry lab I had considered. And the reason I wanted to come here is because I wanted to ship technology that touches people, that is in the hands of people. And Microsoft gives a really good vehicle to do that. Its technology is used everywhere. Anything you develop, if you can do a transfer to the products, your technology will reach out

to so many people. The impact you create is huge, which as an academic is difficult. It's not impossible, but much more difficult. Here it's easier. And yeah, so with nearly all the projects, I have worked very closely with product teams, with the batteries, with virtual Wi-Fi, with white spaces, with farm beats. So going back to your question, I do enjoy technology transfers. I enjoy working with the product teams, the engineers.

They add a realism to your research, where in research, we are allowed to think wild, to think way ahead in the future of things that anyone else, if you tell anyone else, they say it's not possible. You can go target that question. And the engineers help you stay grounded at the same time. That is, yeah, you can come up with a wacky idea, but how do you take it to the real world? It's something which I think being at Microsoft, that's an edge that you get while being here. You're in proximity to so many other researchers. You're in proximity to the engineers, to the sales team who actually talk to customers.

And you go talk to customers as well. So that's one of the strengths, one of the benefits of working at Microsoft Research. I love this so much. I've got a Venn diagram going in my mind of interdisciplinary research efforts, you know, across different fields that you get here. And then you've got the transfer into the product groups. I mean, I've heard actually recently that there's researchers that end up in product groups and I'd call them human tech transfer. Yeah. that there's researchers that end up in product groups and I'd call them human tech transfer. But it's just sort of this dynamic flow of ideas and exchange.

And I love the fact that you're saying, hey, how does this impact a customer even from a research perspective? Yeah, no, but here you get a holistic view. And having been here at Microsoft for 13 years now, it adds a new perspective. Like, for example, with FarmBeats, it's a project and our collaborators are from computer vision, machine learning, hardware, agriculture, audio, networking systems, batteries. You touch every possible research and then you come up with something and then you talk to customers.

You further tweak your research in a way that can then actually transfer to the products as well. You're not just writing papers here at Microsoft. Which is exciting, even though the papers are fun if you're like a writer. Harry Shum refers to what he calls an invisible revolution. It's the system that becomes transparent and you just have an experience with the technology without knowing what's your take on the invisible revolution for technology yeah i think it's absolutely needed and that's something where i'm trying to push on that through my research as well for example with farm bees taking to a farmer who might not even be tech savvy, most of them are not, and helping them benefit from all the latest advances in computer vision, machine learning, networking systems, all of that. Bringing it to the farmer, but to the

farmer, they don't realize all the things in the background that's making it happen. All they see is the benefits of that technology. Doing core research is important. The other part is taking that research in a way where it reaches everyone. It democratizes the technology, right? It gets to everyone out there and everyone can benefit from the advances that you do.

And I think more and more of it will happen. So one of the things I'm seeing here at Microsoft is a breadth of projects that really surprise me. How does this work at Microsoft that you could do something in batteries, you could do something in functional programming languages, you could do something in quantum, you could do something on a farm? Yeah, no, that's a great question. And this is something which is the unique aspect of Microsoft research and Microsoft. And this is what I doubt you'd get it anywhere else.

And a lot of it is driven by our leadership from Eric to Harry to Satya, where everyone's promoting this growth mindset within Microsoft, where you're coming to Microsoft, not just because you're doing a job. You're coming here because it helps you fulfill a passion, something that you're, there is this new culture, this new growth mindset where our executives want us to take risks, want us to do something that has a bigger meaning than just your day job. And this is what I think you can do at Microsoft Research as well. Here we get a lot of freedom. We are able to think big. The people around you are super smart. These are the best people in all the different areas of computer science, electrical engineering, and other fields. That's the encouragement that we get

using which we are able to take risks, take things to a new level. And this is one thing where people who are graduating, I would definitely recommend Microsoft to them about these unique opportunities, which I doubt you'll get anywhere else. You know, this whole podcast is an inspiration to next generation. I mean, I'm literally shaking my head.

What advice, kind of parting thoughts, would you give listeners that are interested in getting into this? What challenges lie ahead in the field that you think the next bright minds might need to tackle, want to tackle? What lines of research are top of mind right now? Yeah, so there's some lines of research which are hot in the research community, some other threads of research which I get excited about. I'll talk about the latter because I want more and more people to work on those problems. And that has been the guiding force behind my thread of research as well about how can you make technology more pervasive, bring the benefits of technology to many more people who are disadvantaged right now. One of the statistics I had given earlier was around half the

world's population is still not connected to the internet. How can we get internet to everyone? Electricity, that's one of the other big problems in our lives right now. So I'm working on batteries and we'll start working on the electric grid as well. How do you get more people online in a clean way? With FarmBeats, we're looking at the agriculture problem. How do we feed the world, bring the benefits of all the technology that we're developing to the rest of the population? Just an advice to the audience listening to this, if you expand the scope of your research

to have an element that touches humans, it gives you a different perspective around your research. Your energy levels don't go down. It gives an amazing amount of satisfaction working on these problems. And this is one thing I have seen even with working with my interns. Initially, a lot of them, all of them brilliant, like the best minds from all the top universities. Initially, all of them were working here, sitting in the lab. But the people, once they go out, once you see them going out and talking to farmers, doing this for like one of the interns. He was from Harvard. He was doing the TV white spaces project. He was so motivated by this work that he went out over a weekend. He drove 1500 miles, stopping every six miles to measure what is the signal

span there. He got pulled over by the cops a few times. We don't make all our interns do this. It's something he did and he just came in. And the reason he was doing this is because of the impact that this technology could create. Since then, once this technology has been made legal, we've connected so many people in rural parts of Africa, India, to the internet using this technology. Same with interns here who worked on farm beats. They just love going to the farm. And a lot of times without even telling me, the excitement that they get by having the research touch so many lives is amazing. And that's one of the advice I would give is it's important to do core research. Let's just go deeper and expand the state with the end goal, that how does it benefit the world? How does it benefit humanity? How does it benefit people?

How can you make technology more accessible to them so that their lives are better? Ranveer Chandra, you are using AI for good and not for evil. And it's been an inspiration talking to you today. Again, this is the SMH interview. Thank you so much for coming in. Thank you. To learn more about Dr. Ranveer Chandra and how AI may be coming to a farm near you,

visit microsoft.com slash research.