I build AI agents to do my chores, mostly because I'm too lazy to do them myself.
Right now, I'm an undergrad (UG'26) pretending to have it all figured out, while somehow writing production AI code at a VC-backed startup. Fake it till you make it, right?
- By day: I politely argue with LLMs until they give me the JSON format I actually asked for.
- By night: I'm grinding on weekend side-projects with friends, fueled entirely by caffeine and the delusion that "it'll definitely work on the first try this time."
- In my free time: I will absolutely ruin your movie-watching experience by pausing the film to explain why the cinematographer chose a specific camera angle. You've been warned.
I like building weird, cool stuff that shouldn't exist yet:
- Making AI agents that actually remember things, so I don't have to explain the context twice.
- Simulating human brain reactions to videos (because running real focus groups involves talking to too many people).
- Figuring out how to make a 300-job application process cost exactly $0.50.
Python • TypeScript • LangGraph • vLLM • FastAPI • Whatever new framework dropped yesterday
If you want to talk about building autonomous agents, why standard RAG is overrated, or just want to aggressively debate the best movie of the 2010s—hit me up. I'm highly critical of my own ideas, so I always appreciate a good roast.