Connor Hyatt

TLDR

My last semester at CMU was the most intense stretch of my life. I scaled Lunon from an idea into a funded startup. At the same time, I swept the awards in two of my hardest courses, winning best project, most technical, and best presentation in a class of 140+ students. Every signal pointed in the same direction: stop sitting on what I know and go build.

Raising Close to $2M in 60 Days

While most of my classmates were focused on finals and job applications, I was running Lunon full speed. Our CEO Dhruv was closing investor and enterprise connections, our COO Ryan was running customer discovery, and I solo built the entire product that got demoed to early clients and investors.

During the process I told no one. No family or friends. I didn't post about it. Every ounce of energy went into building, not seeking validation.

Sweeping 17-313: Foundations of Software Engineering

17-313 is one of CMU's well-known courses with many students in CS, IS, and SWE registering for it. The term project had teams contributing to real open source projects. Our team, Piplup Party, split up so each person could attempt their own contribution to a repository of their choosing. I chose LangChain.

LangChain is the most widely used open source framework for building applications powered by large language models. It has over 100k stars on GitHub.

My PR added over 300 lines of new code and was the only one from our team that got accepted and merged into LangChain's main branch: #33989. So we built our entire presentation around it. The whole thing came together naturally because I was already deep in LangChain every night building Lunon. The course project just happened to overlap with what I was already obsessed with.

Standing in front of a packed lecture hall, I was more nervous than I had been in a long time. But somehow, once I started talking, it turned into the best presentation I have ever given.

After every team finished presenting, all 140+ students and course staff voted on the awards. We took home three out of five:

Best Project. Most Technical. Best Presentation Delivery.

The only two we missed were funniest and most creative. Both went to very deserving classmates.

Our team was inducted into the CMU 17-313 Hall of Fame, which lists the best student groups across all semesters.

Piplup Party team with course instructors Michael Hilton and Chris Timperley Team Piplup Party with course instructors Michael Hilton and Chris Timperley

Winning Design of Artificial Intelligence Products

My other term project win came from Design of Artificial Intelligence Products, taught by Dan Saffer and the great Leon Zhang, who was also an early advisor for Lunon. The course had around 40 students and focused on building AI products that solve real business problems, not just technical demos.

I built an agentic AI system for automating insurance claims processing. The product combined document understanding, multi-step reasoning, and decision automation into a pipeline that could handle claims end-to-end with minimal human intervention. The system reflected everything I had been learning at Lunon about building production-grade AI workflows.

What made the win even more meaningful was the competition. All three of my teammates were graduate AI students, and one of them spent the last decade as a McKinsey consultant. The rest of the class was no exception. CMU's AI programs attract graduate students from top companies and institutions around the world, people who come here specifically to sharpen skills they have already been applying for years. Winning against that caliber of competition confirmed that my hands-on building at Lunon had put me on the same level as people with deep industry and academic backgrounds.

Design of AI Products class photo, I was near the centeral TV, and Leon rocked a holiday hat

The Signal Was Clear

I probably studied less this semester than any other at CMU. But every free moment went into building at Lunon, and that constant exposure to real production AI gave me an unfair advantage in every project without ever cracking open a textbook. The gap between what I was learning by doing and what school was teaching me kept growing wider.

When your side project makes your coursework feel easy, you've outgrown the environment.

I am not some lunatic who actually enjoys competing in term project competitions, and I am nowhere near as book smart as many of my classmates at CMU. But I am a builder, and I know how to drive real impact. These wins were just a natural reflection of that. The competitions happened to be a window where it showed up next to everyone else's output, and the difference was obvious. My other courses this semester did not have competitions, but if they did, I am confident the results would have been the same.

What Comes Next

I am going all in on Lunon. The fundraise gave us the runway to move to San Francisco, and I am recruiting some of the most impressive builders and savants I have come across to join the team. I would be a loser not to take a calculated risk and try to capture this amazing opportunity with my co-founders.