I'm a PhD student at Rice University in the Linna An lab, where I design de novo ligand binders: proteins invented from scratch to grip a chosen small molecule the way a key fits a lock. Nature spent a few billion years evolving its proteins. I get to skip the line — sketching brand-new ones on a computer and then asking the cell to make them real.
What I'm working on now
Most of my days live at the seam between biochemistry and machine learning. The modern protein-design stack feels a little like magic: diffusion models such as RFdiffusion hallucinate a protein backbone around a target molecule, sequence designers like ProteinMPNN / LigandMPNN decide which amino acids should fill that shape, and structure predictors like AlphaFold tell you whether the idea will actually fold before you ever touch a pipette. My lab grew out of pioneering work on small-molecule binders and biosensors, and I'm putting those tools to work building binders against six clinically relevant targets — the kind of molecules that, done right, become diagnostics, sensors, and therapeutics.
The dream: type a molecule into a model, and a few weeks later hold a protein in your hand that grabs it. We're not all the way there — but it's getting eerily close.
How I got here
I did my Honors B.S. in Biochemistry at UT Austin, with an emphasis on synthetic biology and an Elements of Computing certificate — the combination that quietly set up everything since. In the Keatinge-Clay lab I fell for polyketide synthases (PKSs): gorgeous molecular assembly lines that build macrolide antibiotics one chemical step at a time. If you can swap the modules on that assembly line, you can coax it into making molecules it was never meant to. I helped do exactly that — swapping rapamycin modules into the pikromycin synthase to probe the limits of how far this engineering can be pushed, synthesizing 100+ novel polyketides and once leading a small undergrad team to build 36 new synthases in two weeks. That work turned into two publications, including a CRISPR-Cas9 system for module-swapping the pikromycin synthase in its native host.
Rotations & detours
Before settling into the An lab, I rotated widely. In the Kirienko lab I built a Breseq pipeline to track how Pseudomonas aeruginosa evolves resistance to R-pyocins — phage-tail-like "molecular harpoons" bacteria fire at their rivals, and a promising targeted antibiotic for nasty infections like those in cystic fibrosis. In the Han Xiao lab I worked on computational small-molecule library screening and sequence-display revisions. Different systems, same itch: take a complicated biological machine apart, understand it, and try to rebuild it better.
Building things for other people
I co-founded ML4BioChem, a Houston-area seminar series on machine learning in biochemistry, with Profs. Linna An and Cameron Glasscock. In one semester we recruited 13 speakers — including Gabriel Corso (MIT, of Boltz/DiffDock), Kevin Yang (Microsoft Research), and Pranam Chatterjee (Duke) — and grew the mailing list past 300 subscribers. I built the club website too, with a GitHub Actions pipeline that auto-syncs the talk schedule from a Google Sheet every day. (This site is cut from the same cloth: I like things that quietly maintain themselves.)
Off the bench
When I'm not designing proteins, you'll probably find me on a bouldering wall, a badminton or tennis court, or losing a game of blindfold chess very slowly. Through Chi Alpha I've led a small group and helped organize service — shoe cleanings, work for a women's shelter, and cooking for people experiencing homelessness. I care a lot about mentoring, too: I think most "hard" ideas are just well-explained ideas waiting to happen.
- Now: de novo ligand binders + ML, Linna An lab @ Rice
- Before: polyketide synthase engineering @ UT Austin
- Always: teaching, building little tools, and keeping the long arc in the open