I'm Ari, co-founder of LearnTAV. I'm also a high school student, which is the part people usually find surprising when they hear I helped start an AI education company.

This summer I did something outside of LearnTAV: I went to TiE's summer entrepreneurship bootcamp. Four of us came out of it with a company called SecondServe, and a working prototype for reducing food waste. Here's what happened.

What TiE actually is

TiE — The Indus Entrepreneurs — is a global nonprofit that's been running entrepreneurship programs for decades. Its youth arm, TiE Young Entrepreneurs (TYE), teaches high school students to build real startups, not simulations.

The format is deliberate. You're put on a team. You get assigned mentors who are working founders and operators. You spend the program doing customer discovery, building a lightweight business model, iterating on what you find, and then pitching to a panel of investors at the end. Winning teams advance to a global competition against teams from India, Canada, Australia, and other U.S. cities.

The thing I didn't expect: almost none of it is about your idea. It's about whether you actually talked to anyone. Our mentors asked "who did you talk to?" more than they asked anything else.

The problem we picked

Food waste is one of those problems that's so large it stops feeling real.

The numbers, from ReFED's 2026 U.S. Food Waste Report:

70M tons of surplus food in the U.S. in a single year
29% of the food supply — nearly one out of every three bites, thrown out
$380B in value, or roughly $762 per person, per year
51M cars worth of emissions, driven for a year
Under 13% of the food that could have been donated actually was

That last statistic is the whole business. The food isn't spoiled. It isn't unsafe. It's fresh, and it goes in a dumpster anyway — because for a bakery closing at 6pm, throwing food away is simply the cheapest and fastest option available.

We talked to local bakeries and cafés and heard the same thing repeatedly. Nobody wants to waste food. It's just that every alternative takes time they don't have at closing.

The problem isn't that businesses want to waste food. It's that wasting it is the path of least resistance. So we tried to make not wasting it the easier path.

What SecondServe does

SecondServe partners with bakeries, cafés, and restaurants to move their unsold, still-fresh food to people instead of landfills.

Here's the flow:

  1. The business boxes what's left. At close, staff put unsold food into boxes separated by food group and allergen. This mattered more than anything else we designed — it's the difference between "mystery bag" and something a person with a nut allergy can actually buy.
  2. A camera scans the box. It photographs the contents, and SecondServe analyzes the scan against that establishment's own menu. That's the piece we're proudest of. Instead of a vague listing, a customer sees what's actually in the box, described in the restaurant's own menu language.
  3. Customers pick it up at a discount. Fresh food, real price cut, and the business recovers revenue on inventory that was otherwise a total loss.

The insight underneath it: existing surplus-food apps ask the customer to accept uncertainty. We thought the uncertainty was the barrier. Remove it, and you widen the market to everyone who can't gamble on what's inside — people with allergies, dietary restrictions, or a family to feed tonight.

How we built it — and why it's a LearnTAV story

Four of us: me, Avyay Sharma, Yash Khatiwala, and Amartya Katamaraja.

We had weeks, not months. None of us had a professional engineering background. And we still walked into the final pitch with something that ran.

We built the prototype using AI development tools — describing what we wanted in plain English, iterating on what came back, testing it, and fixing it. The scanning and menu-matching logic, the interface, the flow from box to listing: all of it built by high school students who learned to direct AI tools rather than fight them.

This is exactly what we teach at LearnTAV, and SecondServe is the honest version of the pitch. Not "AI makes coding easy" — it doesn't. We got stuck constantly. But the gap between having an idea and having something a judge can click on used to be measured in months of learning syntax. It isn't anymore.

When we pitched, SecondServe placed. But the result mattered less to me than watching four students turn a dumpster problem into a working product in one summer.

Three things I actually learned

1. Talk to people before you build anything.

Our first version had no allergen sorting. One conversation with a parent killed that design. If we'd built for two more weeks before asking, we'd have wasted two weeks.

2. A working prototype beats a beautiful deck.

Every team had slides. Fewer had something judges could open and use. The demo ends arguments a deck can only start.

3. Constraints are a gift.

We had a hard deadline and no budget. That forced us to answer "what's the smallest thing that proves this works?" — a question I now ask about everything.

If you're a student reading this

You do not need a computer science degree to build a real product. You need a problem you actually care about, people willing to tell you you're wrong, and tools that let you build faster than you can lose interest.

That's the entire thesis of LearnTAV — learning takes a village. I'm not teaching from theory. I'm teaching from a summer where four of us built a company from nothing.

Want to build your first tool? Our Vibe Coding 101 course takes you from zero to a working app in four sessions — no coding background required.

Ari Thapar is a co-founder of LearnTAV and a high school student. He built the EcoPrompt Chrome extension and exhibited at STEMfest 2026.

Sources

  • ReFED, Progress on the Plate: 2026 U.S. Food Waste Report — refed.org
  • TiE Young Entrepreneurs — tye-boston.org