"Fate leads the willing and drags along the reluctant."
I started writing code to answer questions my biology textbook left open.
I'm Rikhin. I build things at the intersection of biology and computation. The problems there are hard and a lot of the tools to solve them don't exist yet, so I make them.
Right now I'm at the Broad Institute of MIT and Harvard, working on SPECTRA under Dr. Yasha Ektefaie. At Adaptyv Bio I work on protein screening data: binding curves, expression levels, the messy datasets that never behave the way the textbook ones do. Before that I was at MedARC, decoding fMRI brain scans into interpretable neural representations, work that's now under review at ICML. At Purdue I ran computational drug discovery against the STAT3 protein, docking thousands of compounds from ZINC and ChEMBL against its SH2 and DNA-binding domains. My paper, "agentDX: Evaluating Coding Agents for LLM-driven Bioinformatics," was accepted to BIBM 2026.
On the building side, Auteur is an AI copilot for Adobe After Effects that won the Congressional App Challenge. It has an autonomous mode that reads the documentation, writes ExtendScript, and keeps going until the task is done. Calma is an independent verification lab for AI agents in finance. Linkd is a word chain game that's reached over 60,000 players. Vira is a healthcare hackathon organization I founded, now 20 chapters and more than a thousand students.
What ties it together is the kind of problem I like: the ones you have to understand at a few levels at once, the molecules and the model, the user and the math. That's usually where the interesting work is.