Structural Bioinformatics Scientist — Protein Structure & Drug Discovery

Summary

7 years applying structural bioinformatics to preclinical drug discovery. Implemented AlphaFold2 and ESMFold screening workflows at Relay Therapeutics that produced 3 development candidates. Expert in MD simulation (GROMACS, OpenMM), free energy perturbation (FEP+), and large-scale virtual screening. Comfortable bridging computational and medicinal chemistry teams.

Experience

Senior Computational Chemist — Relay Therapeutics (2021–Present)

• Built an AlphaFold2-based conformational ensemble pipeline for intrinsically disordered targets; identified 2 cryptic pockets validated by cryo-EM.
• Ran FEP+ campaigns on 3 lead series (Schrödinger); 78% hit rate predicting ΔΔG within 1 kcal/mol of experiment.
• Automated virtual screening pipeline (AutoDock-Vina + RDKit ADMET filters) screening 5M compounds/week on AWS GPU fleet.

Postdoctoral Researcher — Bharat Lab, MRC-LMB Cambridge (2019–2021)

• Developed PyRosetta protocol for fitting atomic models into sub-4Å cryo-ET reconstructions; used by 6 labs internationally.
• Solved structure of intact SARS-CoV-2 spike in native virion membrane by cryo-ET + AlphaFold refinement (Nature, 2022).

Selected Publications

• Walsh D. et al. “In situ structure of the SARS-CoV-2 spike in native viral membranes.” Nature, 2022.
• Walsh D. et al. “Cryptic pocket identification in disordered kinases using AlphaFold2 conformational sampling.” J. Med. Chem., 2023.

Code Quality Notes

Python simulation wrappers use pydantic for config validation, preventing silent parameter errors. GROMACS MDPs and Schrödinger input decks version-controlled with parameter sweep history. C++ extensions benchmarked with Google Benchmark; memory safety checked with AddressSanitizer in CI.