hi, I'm
I like finding out how things work.
a high schooler in San Jose · immune cells, code, and a lot of questions
(tap a flower)
where I've been so far
2026 → now
research contributor · remote
I curate published CRISPR screens into FITdb: re-running each through MAGeCK, mapping mouse genes to human, and writing notebooks that check their own work.
Jan – Apr 2026
project lead & biology
Five friends, seven weeks, one question: how do you spot a failing CAR-T batch before it fails? First place in Bioengineering, out of 80+ teams.
Feb – Apr 2026
research intern · remote
Ten weeks of R and bibliometrics on a neural-circuit meta-analysis, presented to faculty from Harvard, Stanford, and USC.
works in progress, taped in
first-author manuscript · in prep for Frontiers in Immunology
Shine Raman light on the cells, train against metabolic ground truth, and let the model say pass, adjust, or flag — before the batch is finished and the money is spent.
Bughouse chess · Python + PyTorch · with a classmate, PhD-mentored
One board for the AI, one for its human partner. It has to guess what you'll do next — and forgive what you just did.
FITdb curation · MAGeCK · Scanpy
Every screen I curate ships with its own QC gates and a SHA-256 of every input, so the result reproduces exactly — or refuses to.
If you're curious about the same things — or completely different ones — write me.