There are three things you can do with your summer. Vacation, work at a big company, or the third path: startups. I chose the third this summer.
It didn’t feel like a break. It was an acceleration. Days and nights blurred into lines of code, whiteboard sketches, hackathons, and startup events. The kind of summer where you measure progress not by how many hours you worked, but by how many code PRs you get in and people you get to know.
By the end, I wasn’t just better at writing code, I had a stronger sense of what it means to build under pressure, a clearer vision of my values and the kind of founder I want to become, and a true belief in the mindset that everything is possible if you truly try.
Here are some of my favorite parts of the summer.
Software at SalesPatriot
The bulk of my summer was spent at SalesPatriot, where I worked on everything that needed to get done, from full stack, to algorithms, to building IKEA furniture. Two of my favorite projects included RankMaster and a Packaging Cost Calculator. But honestly, what made the experience unforgettable wasn’t just the code. It was the work situation. Everyone was under 22, working late nights on our laptops, bouncing ideas across the room. It felt less like a job and more like an episode of Silicon Valley.
This was the view from my desk setup. Really gets you working.
The house we were living in came with its own flavor of chaos. One morning, we’d hired someone to haul away trash. When we forgot to pay him within an hour, instead of knocking… he came back and dumped his entire truck of garbage onto our lawn. We got him paid quickly after that, and rented U-Hauls to take our trash in the future.
Then there was the new hire who decided to spend most of his yearly salary on a Porsche. I found myself tagging along to a Porsche dealership, not your typical onboarding experience.
Meanwhile, our CEO had this project from high school called Boys with the Bus, where he and his friends bought a bus, renovated it, and drove it across the country, racking up 2M followers. That same bus was parked outside all summer, with his friends constantly drifting in and out of the house fixing it up. The neighbors weren’t too happy 😬.
We had a beautiful hike near the house, Mt. Davidson, the tallest point in San Francisco. I loved to clear my mind and think at the top, with the city stretched out below.
Somewhere along the way, I also became the house avocado toast chef. I made it constantly, for myself, for coworkers, for whoever happened to be around. It was a simple thing, but somehow made the house feel more like home.
Now into what I actually did:
RankMaster was a system to score and rank government contract opportunities so companies could focus on the ones most relevant to them. This meant wrestling with unstructured data, inconsistent formats, and edge cases that broke the code (a lot 😂). I had never built anything like it before, and I only had two days to get a working prototype up and running. That kind of time pressure could have been paralyzing, but instead of trying to solve everything at once, I treated it like a puzzle:
Define the end goal: A ranked list that actually reflected reality, not just random scores.
Break the problem into steps: Data ingestion, normalization, scoring logic, ranking, validation.
Set milestones: Clean the raw data first, nail the scoring logic next, validate quickly after.
Focus on function before polish: Make it work first, then make it efficient and elegant. V0 of the algorithm took nearly two minutes per opportunity. By the end of the summer, I got it down to two seconds.
The Packaging Cost Calculator was a very different kind of challenge. Instead of abstract algorithms, this one lived at the intersection of math, UI, and user empathy. The tool helps companies estimate how much it costs to package and ship their products given all the hidden complexities, box sizes, weights, military packaging specifications, that don’t appear until you run a real-world scenario. It forced me to translate raw formulas into a clean, intuitive interface that firms could actually use. Getting the backend math right was one thing; presenting it in a way that felt smooth and reliable was another.
On both projects, I moved constantly between frontend and backend. One day I’d be deep in Python optimization, the next chasing a Svelte rendering bug. I got hands-on with Google Cloud deployment and shipping updates without breaking production, all the gritty parts of real-world engineering that don’t show up in school but matter when customers rely on your software.
I’ve realized building software in the wild means living at the intersection of speed and stability. You’re not just writing code; you’re shipping systems people actually depend on.
YC AI Startup School
From Elon Musk debating AI safety to Sam Altman laying out why conviction matters more than consensus, YC’s AI Startup School was like a dream.
2,500+ builders in one room, hearing directly from the people shaping the future: Sam Altman, Elon Musk, Satya Nadella, Andrew Ng, Andrej Karpathy, Garry Tan, Aravind Srinivas, and more. The energy was electric.
At one point during Startup School, someone casually dropped, “I just hit a PR.” Without thinking, I lit up and asked what feature they’d shipped, assuming they meant a pull request. Turns out… they were talking about the gym. A personal record. It was such an SF moment.
Some of my biggest takeaways:
Live in the future. Build what’s missing.
Garry Tan pointed out that the best startups don’t wait for trends, they create them. That idea stuck with me. Instead of chasing what’s hot, the real opportunity is spotting what’s missing and building it before anyone else realizes it should exist.
Iterate faster than anyone else.
- Sam Altman emphasized that speed of learning is the only moat that matters. Startups don’t win by having perfect ideas; they win by learning faster than competitors and acting on those learnings immediately. Iteration speed compounds, it’s how small teams beat giants.
Velocity > starting point.
- Investors care far less about your pedigree or whether your first version is impressive. What matters is the slope of your progress. A team moving quickly from “okay” to “great” is more valuable than one stuck polishing their “perfect” v1.
Execution beats elegance.
- Andrew Ng drove this home: an ugly prototype in users’ hands is infinitely better than a beautiful design stuck in Figma or your head. Execution creates feedback loops. Elegance can come later.
PMF feels obvious.
- Multiple speakers echoed this: product-market fit isn’t subtle. If you’re still wondering if you have it, you probably don’t. When it’s real, the signs are everywhere, users stick around, demand snowballs, and traction is undeniable.
Breakthrough often comes right after the breaking point.
- Elon Musk told the story of SpaceX’s first three failed rockets. If the fourth hadn’t worked, the company would’ve died. Instead, that was the one that changed everything. It was a reminder that real breakthroughs often come at the edge of failure, you only lose if you stop trying new things and pivoting.
And through all the hype, one theme kept surfacing: the future belongs to those who move fast, but responsibly. It’s not “move fast and break things” anymore. It’s “move fast and build things that matter.” I left those sessions buzzing, not just with ideas, but with frameworks for how to evaluate and prioritize them.
YC Summer Conference
If YC Startup School was about scale, thousands of builders, big names on stage, and big ideas flying around, the YC Summer Conference a few weeks later was the opposite vibe, and I loved that contrast.
Instead of 2,500 people, it was a few hundred, a perfect space to deeply connect. At Startup School, I walked away with frameworks and takeaways. At the Summer Conference, I walked away with friendships. I had time to really talk with people, not just shake hands.
I arrived early, met people in line, and immediately felt that same YC current of ambition and curiosity. Throughout the day, I reconnected with friends from programs like Buildspace and got to spend more time with founders like Selin from Delve and Ethan from CaseFlood.
The talks were phenomenal: • Emmett Shear (Twitch, now Softmax) gave a raw, honest account of launching straight out of school, showing how “not knowing the rules” can actually be your edge. • Han Wang & Hahnbee Lee (Mintlify) walked through their journey from student project to multimillion-dollar company in just three years, super tangible and motivating. • And the YC partners, Harj Taggar, Jared Friedman, and Harshita Arora, hammered home the point: the barrier to starting has never been lower for students.
But the real magic was what happened after. Nobody rushed off. Founders, YC partners, and alumni all stuck around, swapping unpolished stories and real advice. I ended up in a deep conversation with Jared Friedman about how to approach building as a college student. He was genuinely curious about our experiences and feedback, which was exciting to see.
The Summer Conference felt less like a lecture and more like a community, builders figuring it out together. It was inspiring.
Applied AI Cognition Hackathon
One weekend this summer, I found myself in a very unplanned adventure: the Cognition Applied AI Hackathon, judged by Andrej Karpathy and other top people in the field.
I didn’t have a spot on the list. I didn’t even know the venue until the morning of. But I showed up anyway.
I worked with Georg von Manstein, a friend from Buildspace, and Elliot Slusky, a high schooler I’d met 2 days earlier.
We built “Kahoot for politics”, helping politicians understand live constituent sentiment. In just a few hours, we built a website that could:
Launch surveys and gather feedback in real time.
Use AI to classify responses by political leaning and track how opinions shift.
Turn messy data into clean, interactive dashboards you could explore with simple questions.
Generate graphs instantly from plain English queries.
We ended up getting 3rd place, which was insane and unexpected!
The crazy part was how quickly everything had to come together. No time for perfect designs or deep planning. But that intensity created a flow state where the only focus was making the idea real.
I learned how much a small, scrappy team can get done under pressure. And hearing feedback directly from judges like Karpathy, Russell Kaplan, Bill Chen, Hao Sang, and others was surreal.
Walking out, I realized hackathons aren’t just about building projects, they’re about proving to myself that I can take an idea from zero to something real in a few hours.
The hackathon pitch demo was posted here.
Notable Capital’s NextGen Fellowship and Pitch
Another highlight of the summer was being part of the Notable Capital NextGen AI Fellowship. The program brought together a global group of curious, ambitious builders to learn directly from some of the sharpest minds in tech and venture.
One of the most exciting moments came at the first-ever NextGen Pitch Day, where I had the chance to pitch Songlingo, my app to learn languages through music. Walking away with 1st place was surreal, but even more valuable was the feedback from the judges, Jeff Richards, Claire Vo, and James Cham, who pushed me to think bigger about both product and execution.
Beyond the pitch, the fellowship was packed with incredible conversations. We heard raw stories from founders like Garrett Lord (Handshake), Vineet Goel (Parafin), Steven Wang (dub), and Gorkem Yurtseven (fal), as well as insights from Hans Tung, Notable’s managing partner. Each session left me rethinking assumptions and seeing new possibilities for what’s next in AI.
But what really stuck with me was the community itself. The fellows, students, founders, engineers, came from around the world, yet shared the same drive and generosity. By the end, it didn’t just feel like a program, it felt like a network I’ll stay in touch with and keep learning from for years.
Huge thanks to Esther Kim and the entire Notable Capital team for building such a thoughtful, energizing experience.
Outside of the hackathons, late-night coding sessions, and YC events, one of the most unexpectedly meaningful parts of my summer was becoming part of the Jewish community at Chabad SF. I got close with the other interns staying for the summer. Every week, I’d make my pilgrimage on the Muni to Chabad, probably saved hundreds in Ubers this summer :). Between Shabbat dinners, casual hangouts, and conversations that stretched late into the night, it became a grounding counterbalance to the intensity of the tech world.
There’s something about unplugging from code, sharing a meal, swapping stories, and being welcomed into a warm community that recharges you. It gave the summer a rhythm and a pause, and reminded me that while building is exciting, it’s just as important to take breaks and clear your mind.
City of dreams
SF is a city of dreams. You can walk down Market Street and bump into someone you’ve heard about on LinkedIn. Walking around the city, I suddenly heard: “Adam!” — it was Wilson, a friend I knew from Boston. Encounters like that happened constantly. It felt like the boundary between online networks and real life blurred in San Francisco, and that energy made the city itself part of the experience.
That’s the magic of SF: everyone’s building something, everyone’s in motion, and you’re just one coffee chat away from an idea that could change your trajectory. It’s normal to overhear product pitches in line for boba, or walk past a group debating the future of AGI on the sidewalk. The city runs on ambition, but also on openness, people are genuinely down to meet, collaborate, and dream out loud.
It’s not a perfect place, but there’s nowhere else where possibility feels so tangible. This summer was the first time I felt like I wasn’t just learning tech - I was living it.
