About

I'm a Founding Research Engineer at Topos Bio, a drug-discovery AI lab working on the shape-shifting third of the human proteome behind Alzheimer's, Parkinson's, ALS, and many cancers. I joined as employee #2 and the first full-time machine learning hire.

I'm co-first author on Topos-1, an all-atom foundation model for intrinsically disordered proteins. It cuts ensemble error by 43–79% on key experimental benchmarks against leading models including AlphaFold-2 and BioEmu. I built its first working prototypes from scratch and led the multidisciplinary effort across machine learning, computational chemistry, and structural biology through to the full technical report.

I'm now building the next model.

Research
Topos-1 2026

An all-atom foundation model for intrinsically disordered proteins.

Tomas Salgado*, Andre Graubner*, Scott Leishman*, Malhar Kute, Yiran He, Benjamin Rousseau, Connor Bybee, Anthony Giannetti, Xiaotong Lu, Anne Pipathsouk, Jake Drummond, Robert Galemmo, Fernando Martinez, Amir Khosrowshahi, Ryan Zarcone, Grant Rotskoff†

Conformational ensemble generated by Topos-1

Instead of predicting one static structure, Topos-1 generates atomic-resolution conformational ensembles: the full distribution of shapes an intrinsically disordered protein occupies. It outperforms AlphaFold-2, Boltz-2, Chai-1, and BioEmu on key experimental benchmarks, cutting ensemble Rg error by 43–79%, while running about 1,000× faster than molecular dynamics. Its ensembles were successfully used to rank internal small-molecule candidates against a disordered prostate-cancer target in agreement with lab-measured potencies.

Technical report →
Writing
Background
Open source
Oyster: led open-source development at ColorStack, directing 70+ contributors serving 10,000+ underrepresented students
Previously
LinkedIn, Datadog & AWS: Kubernetes infrastructure, cloud integrations & service mesh
Honors
Neo Scholar · Rawlings Cornell Presidential Research Scholar
Education
Cornell University: B.A. Mathematics & Computer Science