Community Spotlight: Sahil Khan

Sahil Khan, a final-year computer science student from India specializing in AI and machine learning, built KCWale, a complete online ordering platform for his brother's cloud kitchen, with an initial investment of just ₹100. The platform, hosted on Render's free tier, initially faced server sleep issues after 15 minutes of inactivity, which Khan solved by using UptimeRobot to periodically ping a /health endpoint, ensuring the backend has not gone to sleep since. Khan's focus on solving real-world problems with AI agents and cost-effective infrastructure highlights the practical application of his studies.

Sahil Khan is a final-year computer science student from India specializing in AI and machine learning. He spends much of his time learning about AI agents and figuring out how to turn ideas into things people can actually use. One of those ideas came from close to home. When Sahil saw how his brother was managing his cloud kitchen, he decided to build a complete online ordering platform for the business. The result was KCWale https://www.kcwale.in/ , built from scratch with an initial investment of just ₹100. We talked with Sahil to talk about building software for a real business while still at university, solving infrastructure problems on an almost-zero budget, and what changes when people start depending on the things you build. Finding problems worth solving Sahil’s interest in computers and AI started at school and eventually led him to study computer science, specializing in AI and ML. Today, he’s particularly interested in building AI agents. But learning new technology on its own isn’t what keeps him interested. Over time, I realized that what I enjoy most isn’t just learning new technology, but looking at problems around me and thinking, ‘Can I build something to solve this? His brother’s cloud kitchen gave him a chance to do exactly that. Instead of approaching it as another college project, Sahil set out to build something his brother could actually use to run the business . Building KCWale for ₹100 Before KCWale, Sahil’s brother was managing the business through WhatsApp and a separate billing system. Sahil built an online platform to handle the whole ordering process instead. Customers can browse the menu, search and filter items, manage their delivery address, choose between online payment and cash on delivery, track orders, and access digital bills and previous orders. On the business side, an admin dashboard gives Sahil’s brother a central place to manage products and users, receive incoming orders, and update their status. Day to day, customers place orders through KCWale, those orders appear in the dashboard, and the kitchen processes them while customers follow their progress online. But Sahil didn’t want his brother to make a large investment before knowing whether customers would actually use the platform. The entire setup initially cost him just ₹100, essentially the cost of the domain. The idea was simple: get customers comfortable ordering online first, and as the business grows, we can invest more into the platform. Keeping the cost that low meant making some compromises on infrastructure. It also gave Sahil another problem to solve. The server keeping the server awake Sahil hosted KCWale’s backend on Render’s free tier. It kept hosting costs down, but there was a catch. The server would go to sleep after 15 minutes without activity . When the next request arrived, the server first had to wake up. For someone trying to place a food order, that could mean an unwanted delay before the site responded. Sahil initially considered writing a script to periodically send a request and keep the backend active. Then he spotted the problem with that solution. I realized I’d need somewhere to run that script 24/7, which would basically create the same problem again.I’d need another server just to keep my server awake. Instead, he started looking for a monitoring service that could send those requests for him and found UptimeRobot. Sahil added a simple /health endpoint to the backend and configured UptimeRobot to periodically send a request to it. The regular requests keep the backend from sitting inactive long enough to go to sleep, without adding another server or increasing the cost of running KCWale. Since setting it up, Sahil says the KCWale backend hasn’t gone to sleep once. When downtime means losing orders The technical problem was straightforward enough. Running software for a real business made its consequences more obvious. A customer opening KCWale during a dinner rush isn’t thinking about hosting tiers, server inactivity, or cold starts. They want food, and they have plenty of other places to order it. When someone visits an online food ordering website, they’re usually there because they’re hungry or craving something.They’re not going to wait around if the website isn’t working, they can simply open another food ordering platform and place their order somewhere else. For a small business, that means even a short availability problem can turn into a lost order and potentially leave a customer with a bad impression of the brand. It changed the way Sahil thought about reliability. Problems that can feel relatively minor during development become much more important when real users are on the other side of them. It was a good reminder that when you build software for real users,things like uptime and reliability aren’t just technical details, they directly affect the business. That’s one of the differences Sahil has found between learning how to develop software and actually shipping it. Sharing the things that work Sahil later shared his KCWale setup publicly https://x.com/terminusio/status/2082683119040401582 , including how he was using UptimeRobot with the Render backend. Part of the reason was simple. He knew other students and developers were likely running into the same problems . Building on a limited budget often means finding practical ways around infrastructure constraints, and a solution that took one developer hours to find might take someone else only a few minutes if it’s already been shared. If I come across a useful trick or find a simple solution to a problem, I’d rather share it than keep it to myself.Something that took me a few hours to figure out might save someone else that time, and that’s valuable. Sahil also sees building in public as a way to take part in the developer community while he’s still learning himself. We’re all learning from each other, and sometimes a small post about a problem you solved can help someone who is stuck on exactly the same thing. Sahil doesn’t claim to have all the answers. He instead shares what he discovers and learns from other people building along the way. Real users teach you what tutorials can’t Sahil’s biggest piece of advice for other students shipping their first production app is not to spend too long preparing to build one. Don’t wait until you feel ready.Build something real and put it in front of people. It’s easy to spend months learning technologies, following tutorials, and building projects that never leave a laptop. Putting software in front of real users introduces a completely different set of lessons. Putting software in front of real users introduces a completely different set of lessons. Suddenly, decisions around deployment, payments, security, and performance have real consequences. Bugs surface in ways they never did during development, and users quickly show you which features actually matter . Sahil recommends keeping that first production app small and the initial costs low. Expensive infrastructure and a long feature list aren’t prerequisites for solving a real problem. Keep the first version small, put it in front of real users, and let the problems you encounter show you what to learn next . Some of my best learning experiences have come from solving problems I didn’t expect to encounter. KCWale has given Sahil plenty of opportunities to do that. What started with the way his brother was managing orders became a live ordering platform used by real customers. Keeping the initial investment close to zero created infrastructure constraints. Solving those constraints taught him more about reliability. And finally, sharing those solutions has given other developers a chance to learn from them, too. For Sahil, that experience sums up the difference between writing code for a project and building something people depend on. A project teaches you how to code; a product teaches you how to build for people. See what Sahil is building Check out KCWale to see Sahil’s online ordering platform in action, or follow his work on X to see what he builds next.