I started college at 16 and I've been dragging half-formed ideas into working things ever since.
First-generation student at USC, studying computer science and business administration. Right now I'm on Core AI at LinkedIn. Before that, a year as an AI research intern at Google and a year and a half as a research fellow at USC.
Most of my work lands somewhere in AI, computer vision, and software. The part I care about most is who gets to do this work at all, so I volunteer with the Viterbi Impact Program teaching STEM around Los Angeles.
Experience
6 tracks
Intelligent motion planning with acceptable contact
A motion planner for robots working in clutter. Instead of treating every object as something to route around, it decides what is safe to touch. A vision-language model assigns each object a semantic cost, those costs feed an anisotropic cost map encoding how safe a push is from a given direction, and a contact-aware A* search plans over motion primitives to find trajectories that permit the contact worth making.
Agent memory
What a system keeps, what it recalls, and when it decides either one matters.