Contributed to development of Keynome Cloud, a cloud platform for bacteria species ID and AMR (anti-microbial resistance) profiling from sequenced DNA. Worked on development of backend data processing pipelines in Python deployed on GCP w/Kubernetes, and on the client frontend in React.

Curated MicrohmDB, dataset of sequenced DNA and AMR profiles for 90K+ bacteria isolates, the largest of its kind. Developed ETL pipelines in Python for cleaning and transforming raw sequencing data and AMR test results (from dozens of hospitals, biobanks, public data sources, etc.) into datasets for use in training ML models for predicting AMR profiles from bacteria genetics.

Trained machine learning and deep learning models to predict battery State-of-Charge and

Trained Bayesian neural networks to model cooling tower dynamics for model-based reinforcement