[embedded content]
For farmers in drought-prone regions, every drop of water counts. But knowing exactly where and when to irrigate crops isn’t always easy, especially for smaller agricultural operations that may not have access to expensive monitoring technology. Two brothers from the University of Texas Rio Grande Valley (UTRGV) believe artificial intelligence could help change that.
I recently had the opportunity to interview Edgar Bello Gonzalez and Eduardo Bello Gonzalez, the winners of Zoho’s first-ever “From the Ground Up” AI Hackathon. Their winning project, Dos Ojos, is an AI-powered crop monitoring system designed to help farmers make smarter irrigation decisions using satellite imagery, drones, and simple text messages.
What I found particularly interesting about their story is that the inspiration behind Dos Ojos comes from personal experience. Growing up in Mexico, the brothers witnessed firsthand how severe water shortages affected their community, damaging crops, threatening livestock, and driving up food prices. After moving to Texas’s Rio Grande Valley, they began noticing similar environmental challenges facing the region’s agricultural community and decided to do something about it.
Their approach combines free Sentinel-2 satellite imagery with optional drone-based monitoring to identify crop stress and recommend where irrigation water may be needed most. Rather than requiring farmers to learn complicated software or navigate technical dashboards, Dos Ojos delivers recommendations through accessible, bilingual SMS messaging.
The brothers also bring complementary skills to the project. Edgar, a computer science major with a minor in electrical engineering, handles much of the technical development. Eduardo, an electrical engineering major with a minor in computer science, focuses on project planning, farmer outreach, and communicating the technology’s potential benefits.
Living in an agricultural community myself, I understand how important water management is to farming operations. What caught my attention about Dos Ojos is its potential to put technology that might otherwise seem out of reach into the hands of smaller farmers without requiring major investments or extensive technical knowledge.
During our conversation, I asked the brothers about their experiences growing up in Mexico, how those experiences inspired Dos Ojos, the technology powering their solution, and what they learned from testing it with local farmers. We also discussed their experience participating in Zoho’s hackathon and what comes next as they work to turn their award-winning prototype into a tool that farmers can use.
Here is my conversation with Edgar and Eduardo Bello Gonzalez about their journey, their award-winning technology, and their vision for helping farmers make better use of limited water resources:
Leland McFarland: Hello, my small business community. My name is Leland McFarland, and I am here with Edgar and Eduardo Bello Gonzalez, who are the winners of Zoho’s From the Ground Up AI Hackathon, and I got a few questions for them. Edgar, Eduardo, glad to have you on. First of all, congratulations on winning Zoho’s From the Ground Up AI Hackathon. Can each of you introduce yourselves and tell us a little bit about what you’re studying at the University of Texas Rio Grande Valley?
Edgar Bello Gonzalez: Yeah, absolutely, I’ll go first. So, my name is Edgar Bello Gonzalez, and at UTRGV, for right now, I’m studying major in computer science and a minor in electrical engineering.
Eduardo Bello Gonzalez: Okay, um, I’m Eduardo Bello Gonzalez. I’m also in UTRGV, but I’m doing electrical engineering with a minor in computer science, so I have it flipped.
Leland McFarland: Awesome. Yeah, I actually that’s my background. I went to Oregon State University. I got my major in computer science. I tried, I dabbled in electrical engineering and then I found out I’m not gonna make any more money, and so I was like, okay, I’ll just focus on computer science without the less hours. But awesome that you guys are going all the way through with it.
Edgar Bello Gonzalez: Thank you.
Eduardo Bello Gonzalez: Yeah, I mean we have fun with it.
Edgar Bello Gonzalez: Yeah. I always love to meet a fellow computer science.
Leland McFarland: Yeah, it’s you know, I love seeing it out in the, you know, out in the wild as well. All right, so you guys are brothers as well as teammates. What was it like working together on Dos Ojos? And how did you divide up the responsibilities between the two of you?
Eduardo Bello Gonzalez: Well, I mean, I guess, I mean, we’re really used to working together. We, like we’ve been together our whole lives. We’re brothers, we’re twins actually, we’re fraternal twins. So, I mean, I guess he’s like the most reliable teammate I’ve ever had, to be honest. So it just makes everything easier. And then, I mean, we were both like really committed to the project, so like we both knew we could account for each other doing their part. To be honest, Edgar did most of the, like, the technical part, most to not say all of it. Because, you know, he’s computer science. He’s one of the best people I know, like, doing computer science. Like he would, he would explain to me what he’s doing, and I, and I do understand it, but I could never do what he’s doing. So he was taking, like, responsibility of that, and I was doing more of, like, doing, like working on the presentation, working on the speech, learning how to sell our idea, contacting the farmers. So basically I was doing all of the logistics behind it, and also like trying to give feedback, ideas on how to make it better, what we can implement, and how to, like, make it grow even more while he was working on, like, the actual, the actual computer science side of it.
Leland McFarland: Nice. How about you, Edgar?
Edgar Bello Gonzalez: Yeah, well, what Eduardo said is pretty much it. The tasks were mainly divided into, like, the project, like the program itself, and everything else. So, I pretty much took in charge, was in charge of everything that had to do with the project. Also, I wanted to explain to Eduardo like how it works so that way, on the presentation, any of us could answer any questions that the judges would make. And, yeah, he’s, he’s like really, really good at organizing everything. Even like some of the judges told us that something that we, that we decided like what the budget was going to go for, or like what was like the plan that we had ahead, was like really good, really complete. And, and that was pretty much all Eduardo, because he took care of everything, pretty much almost everything else that was not the program itself.
Leland McFarland: Sounds like you guys had some real good synergy between the two of you. All right. So, you two grew up in Mexico and experienced periods when your city didn’t have any water. How did seeing that, the impact of the water shortages on farmers and local communities, influence the idea behind Dos Ojos?
Eduardo Bello Gonzalez: Actually, you can go ahead. I think this one’s this one’s yours.
Edgar Bello Gonzalez: Okay, okay. So, yeah, we did grow up in Mexico, and there was quite a while, like there was a really long time where we didn’t have a lot of water. Sometimes we had to to wait like at a certain time of the day to shower, because maybe like during most of the day there was no shower. There was only water like during the morning or maybe on the evening. And for, it was personally bad for us, because not having that much water is, it’s really bad, but not only for us, also for the market itself, because the crops were like dying, and that also meant that the cattle was dying. So, things were starting to get more expensive, because there was less things. So, it was just like a lot of things that that led to each other because there was like no water. So, when we got here, which is like really, really close to where we were living. Oh, yeah, to where we were living. It had the very, very similar conditions, because we actually live like 2 hours away from where we used to live. So, the conditions are pretty similar. And we were starting to see that the patterns were like kind of starting, like how we looked back on Mexico. So, we didn’t want that to happen where we live now like how it happened where we lived back in Mexico. So, we decided to take, like take matters into our hands and thought, okay, how can we, how can we try to solve this problem? And that’s when we, oh, my bad, and that’s when we came up with this idea that that utilizing the water the most efficient way will definitely help us like save the crops and therefore save the cattle as well. And also because here on the RGV, there’s a lot of crops, but like a lot, even more than where we used to live.
Eduardo Bello Gonzalez: I mean, so, I mean, I guess when we first looked at the problem, like, of course, it was like, oh, there’s lack of water, like maybe like sometimes we want to shower and we can’t, like stuff like that that are more of an inconvenience, because like back then we were in middle school, high school, so it was more of, yeah, like they were inconveniences, because Mexico was going through like a certain drought like 5 years ago, no, like 4 years ago. So, it was like, it was, it was an inconvenience, yeah. But then, like, of course, when we moved here, like he was saying, a lot of the, like, climate and droughts and all that were similar, but we were like, okay, of course we’re not going to solve the whole problem of saying, okay, now we’re going to save water and we’re going to get like all of it and make sure that there’s no droughts ever. So, of course we had to like focus our idea on something. So, I mean, as he was saying, like the RGV, it has a lot of growth, like of growers of crops, of like agriculture. So, it was kind of like the, like the best target audience that we could find in the valley. And then, like, it’s such a core activity both in Mexico and in the RGV, so it was like everything started with the, like the idea in general of rationing water. But of course, like, we wanted to go into something more specific, something that would target something more like specific for the valley. So, that’s why it kind of like shifted of course, going more into agriculture, growers, water distribution for crops, and all that.
Leland McFarland: Great. All right, so after moving to Rio Grand Valley, you began seeing, you know, as you said, you began seeing similar drought and watering conditions. Was there a particular moment after you moved that you realized that this was the problem that you wanted to solve? Was there something like, was that that light bulb moment?
Eduardo Bello Gonzalez: I mean, I guess it’s just like, because since the moment we moved, we realized that agriculture here was like way more important than we thought. So, like when we moved, we realized like all of the, like agricultural departments they have and like how important, like agriculture and growing is here. And we kept going back and forth, like we keep going back and forth to Mexico, we go every weekend. So, like the problem, maybe like of course right now it’s not as present as it was like that many years ago, because they’ve kind of been working on it, but sometimes there’s still the occasion in which, in which there’s no water, like in specific places. So, the issue is still permanent. So, it’s kind of like having, having that situation of, since we’re going back and forth, we keep seeing it, like both here and there. But now we have like this insight of, like we realized that it’s, this is like way too present here. And then, of course, like since we’re now here, like we keep seeing more news and we learn more about the place and we like go around the city. So, we realized those kinds of things. It’s more of a, like a combination of the, of the new environment combined with the, with like the present and current issue that we’ve noticed. Like at some point, they were just like, oh, they just keep running in our head until, like at some point, we were asked if we could help, how we’d do it, and it just kind of like popped out of thin air based on what we’ve seen like all this time.
Edgar Bello Gonzalez: Yeah, for me, I wouldn’t say that there was a specific moment. It was just something that has always been on the background since the time that we moved here. Every time that, well, almost all the time that we looked at the, at the grass, it looked kind of pale instead of green. And when I saw like the prompt of, of like Zoho Hackathon, it immediately came to my mind, like, yeah, I’ve seen a lot of dryness here on the RGV, like throughout all the time that I’ve been here for like the last 2 years. So, it’s not like I realized like there was, like it was going to happen the same as Mexico. It’s just like since the time I came here, I’ve seen that it’s been getting slightly drier and drier, and when I saw the topic of the hackathon, I thought, okay, this might be a good opportunity to do something about it.
Leland McFarland: Great. All right, so for someone who isn’t familiar with agricultural technology, can you explain what Dos Ojos does and how a farmer would actually use it?
Edgar Bello Gonzalez: Okay, yeah, absolutely. So, how Dos Ojos works is that, okay, so it is divided into three parts. It’s the satellite imagery, the drone imagery, and the and the SMS message, which is the way that the farmers contact the the actual program itself. So, I’ll start with the SMS, which I think is the simplest one. They can just send a message to a bot that we have that just gets answers and gives replies in order to get information from the from the farmers about their farm. And at some point, it asks you, it asks you to get like, okay, where is your farm located? And then it asks you where your farm is located. You can open it, you can send the bot a link on Google Maps, and then you can select the area of your of your crops. Okay, at that time, that’s when the satellite imagery comes in. And it uses free satellite data from a satellite that it’s from AWS, it’s called Sentinel-2. And it goes around the Earth, I would say like every 5 days or so. So, it just, whenever you ask for a specific part, it just fetches in real time the information from that farm, like the one that is closest to it, like let’s say the satellite passed your crops 2 days ago, it will just get the most recent one. Not only that, it also gets like a few days and months ago in order to compare it on like how it looks now to how it looks, or how it used to look before. And from the satellite, since it’s like really far, there’s not a lot of things that you can see, but the few things that you can see are like really important. The thing that you can see from like really, really far is how green it looks. And how green a farm looks tells you almost all you need to know about it. So, it takes that into account, and it can also compare it to previous months, previous years to see if it used to look greener or has it always been like this. And that’s pretty much what it takes. It takes into account how green it looks versus how green it’s been looking throughout the years. And then the satellite part, I mean, the drone part is the second eye, which is, I want to say, slightly more accurate because it’s closer to you and it’s just a few feet on the air. And this one can get information from like each specific crop that is planted. Like, using computer vision, it can count like all the crops that there are, and then like box them into each itself, and it can also flag them like they’re healthy, you should look out for them, or they’re like dead. It also does this by comparing like each crop to the ones next to it. But the only problem with this is that if all of them are like kind of dying, they would all look the same, and you might think that they’re healthy, but the satellite can tell you otherwise, because it can look like exactly how it looks and how it’s supposed to look. So, for the farmer’s point of view, you just have to give it like the satellite data, like where is it located and select it. And then, if you have the chance, because you can also select only satellite or satellite and drone, you do not need to have the drone to do it. You can use satellite only as well, and it will give you like a really good estimate as well. But, of course, it will be even more precise with the drone as well. And the farmers do not need to know at all how this works. They just need to know information about their farm, like what they’re growing, where is it located, and when did they irrigate it.
Eduardo Bello Gonzalez: I mean, so, I guess something we told a lot is that, I mean, of course the program is meant to be used with the satellite and the drone, but if like if anyone were to want to start just with the satellite, they’re completely free to just do it, try it out for a while, and then see how, like what kind of recommendations and information they get. And then, once they get used to it, they can try it and get a drone and gather pictures too. So, it’s kind of you don’t have to go all in since the beginning. You can start with one eye and then add the other one. And then, something else is that the program, it gives you recommendations for water usage, but it also tells you a lot of information about why. So, also, like what we used the program, like they just send a text to the, because we’re working on getting servers. For the moment, it can only run if he has the laptop open, but once we get the servers running, like they would just have to send an SMS message to like to a chatbot and then just tell them like, ‘Oh, my field is located here. I’m growing this. I watered it like this time ago,’ like just some basic information about the field, and then that would be pretty much it. The chatbot is going to tell them recommendations about how to look for the crops, and it will also give them access to a file of the background of why it’s giving them all of those recommendations. So, the farmer can choose if they only want the recommendations or if they want to see all of the background information, too. So, it gives a lot of freedom of how they want to use it, how involved they want to get with the usage of it. So, yeah.
Leland McFarland: Nice. All right, so where does AI come in all this whole process? How does it help identify crop stress and determine where irrigation water may need to be applied the most?
Edgar Bello Gonzalez: Okay, no, that’s a great question. So, AI specific, in general, there’s like three sections for AI. There’s for the SMS message, there’s an AI that is reading the comments, and then it also gets all the information from the satellite, all the information from the drone, and all the information from the farmer. And with all that information, it takes into account to make a recommendation for it. But also, something really important is that it’s, well, at least that part of the AI, it’s a language model. So, it can be like, sometimes it may be wrong and it thinks that it’s correct. But we solve this by, just before like sending the text, we make sure that it actually matches to the soil data related with the irrigation data. If it matches, then that means that the recommendation is like true and it’s not hallucinating. But if it’s not matching what it should be, what the numbers are saying, then we tell it to like redo it again, like it did something wrong. And for the other section of AI, there is the computer vision, which is three different neural networks, all trained using only NumPy, all trained like separately as well. So, there’s one that counts like how many trees there, not trees, but like how many crops there are. There’s another one that gets the height of the of each crop, because it’s also, that also tells you like how healthy it is. It’s really important, like how the health related to the color of the crop. And then the other one, it’s what I mentioned before that compares, compares each one of them to the ones that are close to it. And if if one of them is specifically looking like weaker, then it flags it. But then if multiple are looking weaker, it might not flag them, and then, as I said before, it might think that they’re fine, but they might not be fine, and the satellite will flag it then, because it can see it that it’s actually not fine. And also, it takes input from the, something that I wanted to mention earlier was that it takes also comments from the farmer. So, if you, for example, harvested like just half of the field, you can tell that to the bot and the AI will take it into account to make the recommendation, because if you harvest only half of it, then of course half of the crops will look like really healthy and the ones would look like not healthy at all, because they were just harvested. But the model, the language model will take it into account to know, okay, I won’t flag it, because now you told me that they’re actually not, I mean, like they’re healthy, they’re just harvested.
Leland McFarland: Great. All right, so you kind of mentioned this before, but instead of requiring farmers to use complicated dashboards or apps, Dos Ojos can use bilingual recommendations through text messaging. Why was text messages messaging such an important part of your design?
Eduardo Bello Gonzalez: It’s easy to access. It’s easy to use. And SMS, like you don’t need hotspot, like you just need a service. So, I mean, it’s something really easy to access and really easy to use. So, that’s pretty much why. It’s reliable and accessible.
Edgar Bello Gonzalez: Yeah, mainly we know that most of the farmers are not like really up to date with technology. So, we wanted to make it the most user-friendly as possible, like as most as possible. And actually, we found out that it worked, because the most difficult part for using this program was using Google Maps. So, like all the farmers, well, not all of them, but all of the ones that struggle, they struggle with Google Maps, like not even with the bot itself. It was just like getting the location on Google Maps and sending it to the bot. Like, that was the most difficult part. And that’s why we chose to make it SMS, to make it just as user-friendly as possible.
Eduardo Bello Gonzalez: Yeah, because before we went to the competition, we showed the, we tested a program with growers like in Texas. So, those were the feedback, like the hardest part is locating in Google Maps the field. And then it’s also the fact that a lot of growers, like local growers, they manage their operations from the phone. Like they’re just going to be talking to other people about like what to do. So, like their phone is their main line of communication and how they communicate what’s going to happen. So, if they have it, they can just send a text and get more information.
Leland McFarland: Great. They need to upgrade to Starlink and get internet wherever they’re roaming, right?
Eduardo Bello Gonzalez: Yeah.
Edgar Bello Gonzalez: Would be nice.
Leland McFarland: So, what might a farmer actually receive in one of those messages? What information would you provide so that they can make the useful decision in the field?
Edgar Bello Gonzalez: So, it gives like a, well, at the end goal is giving a recommendation. But also, something really important is that it’s just a recommendation. The bot, like the AI, it’s not deciding for the farmer. It’s just telling it like what it thinks it might be the most efficient way to to irrigate, like when to do it and how to do it. But it doesn’t like tell, like it doesn’t require the farmer to do it. It’s just a recommendation. And they’re like free to to listen to it or like listen to like half of what they’re saying. Maybe because maybe they know that they tried something similar and it didn’t work, and maybe they know a method that is better. But anyways, it’s nice to know like, when should I do it? But use just like a different method from the farmer. So, it’s just, it’s just a recommendation.
Leland McFarland: Okay. So, what made you decide to enter Zoho’s From the Ground Up Hackathon, and how much of Dos Ojos existed before the competition even began?
Edgar Bello Gonzalez: So, what made, so, I was, something really funny is that I was trying, I was applying for internships, and then I I really got tired of applying to internships. So, I started looking at the Discord server of my university, and I saw that the chair department of computer science sent an announcement saying that this national AI hackathon was going on and you should apply, you should attend this webinar if any questions. So, that was, I literally saw that the same day that the webinar was. So, I was like, oh, yeah, this looks really interesting. I’m going to get in. So, I started, and then I started making some questions about like, how did it work? What did I need to do? And after that, they were really, well, the Zoho professionals were really friendly, and they solved all my doubts. So, I really wanted to compete after that webinar. And something really funny as well is that when applying for the hackathon, I really took inspiration from the research that I’m doing, because I work as a research assistant. And I came up with this idea, and then I submitted it. And I really thought that we were not going to get it. But then, but then like 2 weeks later, we got the email saying, ‘Hey, congrats on being the finalist. Now, would you want to come to California and present it?’ And something really funny is that, for this point, I hadn’t told Eduardo about the competition. Like, I just signed both of us, because I knew, I knew that if we were selected for any reason, he would totally be down to help me with the project. So, that same day, like not even 1 hour later, I went with Eduardo and told him, ‘Hey, would you like to go to California?’ And he said, ‘Absolutely.’ So, that’s the story of how we applied. And then we just started to lock in for the project. We divided the tasks, and we started working really hard.
Leland McFarland: All right, so now that you’ve won, what happens next for Dos Ojos? What needs to happen to move it from prototype into something farmers can actually begin using?
Eduardo Bello Gonzalez: Well, actually, this last week we met with a, oh, what’s up?
Leland McFarland: Well, I was going to say, you’ve actually talked with farmers, so I mean, you’re you’re already in that development cycle, I’m guessing.
Eduardo Bello Gonzalez: Yeah, this last week we spoke with a, it’s like a connection that the university have with the United States Department of Agriculture. So, basically, like we were seeing if we could get like funding for the project. So, for the moment, like she was just telling us like, ‘Okay, how can I help you? How can we make this, like make this project grow?’ So, we were like discussing, since we already had the budget, because like it was part of our presentation, we’re like, ‘Oh, okay, we can get started with this and like where it is, where it is.’ So, basically, like currently they’re like, well, we’re kind of like in the works to see if we can get like the first steps, which would be like getting the server, basically, just to see if we can get that done. Then we also agreed that, because right now we’re having exams, we’re in midterm season, so we agreed that once we’re over with exams, we’re going to, we’re going to revisit like the, like what’s going next. But basically, we, like the last thing we spoke was they’re going to connect us with even more farmers so they can, like the farmers that are already like associated with the Department of Agriculture, just to see to have them try too, and start getting like, start having them use like the satellite, and then try to see like what we can do about like getting what’s needed to be able to add the second eye and have them try the drone, too. So, yeah, we’re already in the in the works to see how we can make it grow, probably once we’re done with exams and the homework we’re having right now, because university is still going. But, yeah, so we’re trying to make it get to more farmers, and we’re seeing like trying to help them escalate from using the satellite, and then after a couple of times, then go ahead and use the like the drones, too. And then we’re also, we’re discussing like, okay, what other things we could try to program to like to do. So, what we’re discussing with, her name is Ruby, we’re discussing with her like what we could do maybe like try to detect plagues or try to detect like specific kind of worm. So, we would have to investigate like what kind of signatures they leave in the, like in the crops or environment. So, yeah, it’s kind of, we’re still working on on the project, we’re still working on making it grow, trying to get to more farmers, trying to to get people to know about it and trust it, like by using it and seeing the recommendations that it gives. So, yeah, I mean, we’re just keep, like we keep doing it and we can try to make it grow.
Leland McFarland: Awesome. Sounds like a really awesome project. I live in a farming community, too, and I know a lot of farmers that would love to be able to have this kind of tech on hand and ready to go. Thank you for coming on. That’s the questions that I have for you. Appreciate your time.
Eduardo Bello Gonzalez: Yeah, no, thank you for having us over.
Edgar Bello Gonzalez: Yeah, thank you so much. It was a pleasure.
Leland McFarland: All right, and for those watching, I’ll have a link in the description where you can learn a little bit more about these guys’ projects. All right, well, thank you.
Edgar Bello Gonzalez: Thank you so much.
After speaking with Edgar and Eduardo Bello Gonzalez, what stood out to me most wasn’t simply the technology behind Dos Ojos, but the practical problem they are trying to solve and their commitment to making that technology accessible to the people who need it.
We often hear about artificial intelligence transforming industries, but many of those conversations focus on large companies with substantial budgets and technical resources. What I find encouraging about Dos Ojos is that the brothers are approaching AI from a different direction: finding ways to help smaller agricultural operations benefit from technology without requiring them to invest in expensive systems or learn complicated software.
I was particularly interested in their decision to use text messaging as the primary way farmers interact with the system. It’s a reminder that innovation doesn’t always mean introducing something more complicated. Sometimes the most effective approach is to take advanced technology and make it as simple as possible for the end user.
Another aspect of their project that I appreciate is that Dos Ojos isn’t designed to replace a farmer’s knowledge or experience. As Edgar explained during our conversation, the system provides recommendations, but the farmer ultimately decides what actions to take. That distinction is important, particularly in agriculture, where local knowledge and years of hands-on experience can be just as valuable as the latest technology.
Of course, winning Zoho’s From the Ground Up AI Hackathon is only the beginning. The brothers still have work ahead of them, including securing funding, establishing server infrastructure, expanding testing with farmers, and refining the system based on real-world feedback. They’re also exploring the possibility of eventually expanding Dos Ojos to help identify agricultural pests and other threats to crops.
Living in a farming community myself, I can see the potential value of a tool like this. Water availability, crop health, and operating costs are real concerns for agricultural businesses, and having access to additional information could help farmers make more informed decisions.
What makes Edgar and Eduardo’s story especially compelling to me is how they took a problem they experienced growing up and turned it into an opportunity to help others. Their project is still in its early stages, and its long-term success remains to be seen, but they’ve already demonstrated how emerging technology can be applied to challenges that affect small businesses and local communities.
I’ll be interested to see how Dos Ojos develops as the brothers continue working with farmers and expanding the capabilities of their technology. Their journey is a good example of how meaningful innovation can start with a simple question: How can we use the tools available to us to solve a problem we’ve seen firsthand?
