summer 2026. we just packed our bags and booked a one-sided flight to bangalore. no return ticket. no real plan. barely any money. just a half-baked google form we'd filled out weeks ago for something called Uniform2Unicorn at polaris school of technology. typed something decent, hit submit, moved on. then they called and said we won.
so we left. me, tanish, and aditya - three guys who thought we'd figure it out once we landed. and somehow, we did.
we reached bangalore at night, and there was already a car waiting for us. which immediately made us feel like we were more important than we probably were. polaris had offered us four days of accommodation on campus, which meant four days of not worrying about logistics and just focusing on the actual work. we pitched our ideas, sat through interviews, met people, shot some reels, and tried our best to look like we had everything figured out. mostly it worked. there's a kind of confidence that comes from genuinely believing in what you're building, and we had that at least.
around the same time, we were also invited to a pitching event by Masters Union. originally scheduled for the 13th but rescheduled to the 17th specifically for us, which we may have mentioned to each other more than once because it felt nice.
after the polaris accommodation ended, the trip shifted into a different gear entirely. no more campus, no more free rooms. just us, three hotels across the next few weeks, and the city.
the hackathons started around june 19th. we did three of them back to back: AIBOOMI, Reactor.inc, and the Agent Arena Bangalore Hackathon 2026 - an AMD × AICamp × Google collab that had a ₹50,000 first prize.
we won aiboomi and agent arena. for both of them we were building on top of something i'd put together mid-hackathon called Prometheus- a framework that became the base we kept stacking newer stuff on top of, kind of like the world's most sleep-deprived lego set. if you're curious: github.com/AdityaBhatia-agentperry007/promethues. yes the spelling is wrong. no we haven't fixed it.
somewhere in between all of this - around june 23rd or 24th - we started going to networking events hosted by cedat, and that's where things got genuinely interesting.
we met Pratham Pengoriathere. founder of jee simplified and solve arena on youtube. aditya has been watching this guy since 9th grade, which made the whole thing feel slightly unreal. the way it does when someone who existed only on a screen suddenly has a coffee cup in front of them and is asking you what you're building. we sat with him at Tim Hortons in HSR Layout for about two to three hours, talking about education, about the gaps in what exists for jee and neet students, about where technology is actually going in indian edtech versus where people thinkit's going. by the end of it, we had an unofficial deal on the table: build a specific piece of software and he'd pay ₹2 lakhs to use it in his platform.
but the more important thing he said was almost a throwaway line near the end.
“there's this guy named ravneet singh. he was director of technology at vedantu, co-founder. physics wallah offered him their cto position - he turned it down. now he's started a new thing called prolearn and raised ₹30 crore in pre-seed. just on an idea. his office is actually right around the corner.”
and then pratham left.
we sat there for about thirty seconds. then aditya opened google maps.
it took maybe four minutes to find the office. we walked over, went upstairs, and told the security guard we were there to meet ravneet singh. the guard looked at us: three guys in their late teens, no appointment, no context, no credentials beyond vibes. he was visibly trying to figure out what was happening. honestly, fair.
and then ravneet walked past the corridor at that exact moment.
he stopped. he looked at us. there was a pause that felt longer than it was. his expression was somewhere between “who are these people” and “i'm genuinely curious what they want”. and i think the second one won, because he said come in.
we went inside, sat down, and spent the next hour or so talking about everything. who we were, what we were working on, why we'd come to bangalore, and what we were hoping to get out of it. we talked about orca, which was the project we were pitching at the time, an agentic android control thing. we asked about internships, about lors for college applications. and ravneet was the kind of person who doesn't soften things. he told us exactly what he thought, pointed out the real weaknesses in our approach, and talked about edtech with the kind of fluency that only comes from having actually built things at scale. it wasn't comfortable, but it was the most useful hour we'd had in weeks.
the software idea we'd been loosely discussing with pratham - Vivacity- was still pretty rough at that point. ravneet's perspective was what actually shaped it into something real.
here's what vivacity is: you know 3Blue1Brown - grant sanderson, the mathematician who made youtube feel like a film studio? he built Manim, a mathematical animation library that's behind all of his videos. what we asked ourselves was: what if you took that visual language and pointed it specifically at jee and neet students? not at general physics concepts, not at broad topics. at the actual problems. the specific questions students get stuck on where they understand the theory fine but can't see the solution.
there are simulation tools out there - phet, simphy - but they're built to explain concepts in general. there's nothing built around indian competitive exam problems specifically. nothing that takes your doubt in a particular question and shows you the solution animated, step by step, the way your brain actually wants to process it. that's the gap. that's vivacity. we took the design essence of 3b1b's approach, trained a small open-source model on a jee-relevant dataset, and built an engine that generates those animated explanations. the existing forks of manim that other organizations have tried either serve completely different use cases or are basically unusable in practice, so we had to build the pipeline ourselves.
those were the demos we eventually went back to show ravneet.
after that first meeting we tried visiting a few more times. he was unavailable most of those, which is the respectful way of saying we showed up unannounced to someone's office multiple times like this was normal behavior. around july 3rd or 4thhe was actually there and agreed to see us again. we went in, showed him what we'd built, and he was genuinely impressed. we talked for another two to three hours, met a lot of the people working there, and came out with a ₹60k internship. we haven't accepted it yet. but it's there.
there were other people too. Prateek Behera, co-founder of gradcapital - we had a personal one-on-one with him, talked through vivacity, and we're pending approval on a ₹6 lakhs deal. Abhishek Sensharma, founder of lucidml, who we got to know through gradcapital. we spent time at Forge Residency and InnovationNoveltyHQ. we knew founders there personally, which is how we got in, and those late-night rooms with whiteboards and bad wi-fi were some of the best parts of the whole trip.
and then there were two days where we were completely, entirely, not-a-single-rupee broke, camped outside a Shell petrol pump.
i want to be clear about the specific humour of this situation: we had just won ₹50,000 at a hackathon. people were discussing six-figure deals with us over coffee. and we were sitting outside a petrol pump with our laptops, rationing whatever snacks we had, waiting for a transfer to come through. the city had not been informed that we were briefly insolvent. we kept that professional.
it lasted two days. we survived. we moved into the third hotel.
by the time we actually left bangalore - and yes, we eventually booked the return tickets - we'd won two hackathons, built a product that serious people in indian edtech took seriously, met founders and investors across three weeks of barely-planned persistence, walked cold into an office and came out with an internship, locked a ₹6 lakhs deal, and spent two days broke at a petrol pump in a city that had no idea. we came with nothing and left with more than we could have planned for.
we came for a competition. the competition was just the door.
Vivacity is still being built. What comes next - that's what we're figuring out.