Hey, I’m Beckett. I’m a student at Stanford University. I love building things that address inefficiencies in society to hopefully make the world a better place.
I co-founded Realer Estate, a platform helping 140,000+ New Yorkers find rent-stabilized and undervalued apartments. Our work has been featured in The New York Times, CBS News, and ABC7, and we raised $100k.
More recently, I’ve been spending a lot of time researching how AI inference can be made dramatically cheaper and more energy efficient by performing computation directly inside memory chips. Specifically, I’m investigating how commercially available commodity flash can be used for neural-network computation at trillion parameter scale.
Housing and inference hardware are very different fields, but I think about both of them with the same underlying curiosity: how can this (unnecessarily) inaccessible resource be made more abundant for everyone?
I’m interested in economics, physics, history, computing, and building things that don’t exist yet. I also love to read and work on experimental projects on the side.
Work
Realer Estate
A housing search platform helping 140,000+ New Yorkers find rent-stabilized and undervalued apartments.
Compute-in-Memory
Compute-in-memory inference chips to run frontier AI models on flash memory, making AI inference hardware significantly cheaper (starting with Kimi K3). More on this soon.
Press
- New York Lacked an Affordable Housing Portal. So These Teenagers Made One.
- NYC teens’ website designed to tackle affordable housing crisis
- Two Brooklyn teens saw a broken system and built a tool thousands are using
- Teens create free tool to find affordable housing in New York
- Teens launch portal to help New Yorkers find affordable housing