Software Engineer Intern · Remote · Aug 2024 – Jul 2025

Stack: PyTorch, FastAPI, React

Gene2Lead applies machine learning to drug discovery. I spent a year working on the models, the APIs that serve them, and a tool for understanding what the models learn.

Deep learning on protein-drug interactions

I implemented featurization, GPU training, and inference in PyTorch. The models analyze 7 million protein-drug interactions, which speeds up site selection in drug discovery.

Candidate selection APIs

I deployed candidate selection REST APIs on AWS SageMaker. More than 30 commercial users now obtain a ranked list of candidates, scored on reactivity, programmatically.

Seeing inside the black box

I built a proof-of-concept visualization tool with React and FastAPI. It helps people understand proteomic data and build trust in black-box ML models that run in production.