Ph.D. Candidate · Computer Science · UMass Amherst / Amherst, MA, USA
Computational biology · ML for genomics
Machine learning for the biology of antibiotic resistance
Ph.D. candidate studying drug resistance in Mycobacterium tuberculosis. Advised by Prof. Anna Green, SAGE Lab.
Open to postdoctoral & research scientist roles · flexible start, 2027
Why it mattersIn 2023, an estimated 410,000 people developed drug-resistant tuberculosis, and only 43% were diagnosed and started on appropriate treatment. WHO, 2024
Sequence-only ML models in genomics look impressive on benchmarks and then fail on new lineages, new genes, new populations. I build models that stay accurate and explainable by injecting the biology we already know — protein structure, evolutionary constraint, multi-omic context — into the model itself.
Research highlights
What I'm working on
Benchmark
BIG-TB
A unified dataset and evaluation framework spanning 17,942 isolates and 11 drugs. I led the protein-side benchmark, built to compare resistance-prediction models on equal footing, including on lineages they never saw in training.
Forecasting
FARM
Forecasting which uncertain variants will later be reclassified as resistance-causing, by combining 3D structure, Rosetta energetics, and protein-language-model features, then scoring 4,525 variants for follow-up.
Structure-aware models
Fused Ridge & 3D clustering
Two published papers on putting protein structure into the model. A structure-regularized linear model (ICLR MLGenX 2025) reached a mean AUC of 0.766 across nine genes, ahead of plain ridge (0.755) and zero-shot ESM-2 (0.603). An eLife 2025 classifier showed that 3D proximity predicts resistance mutations better than sequence distance (F1 94.6% vs 92.8%).
Selected publications
Recent papers
- 2026
bioRxiv Preprint · under review at PNAS
FARM: Forecasting Antibiotic Resistance in Mycobacterium tuberculosis Using Biophysics and Machine Learning
- 2026
bioRxiv Preprint · under revision
BIG-TB: A Benchmark for Prediction and Interpretability of Sequence-Based Machine Learning Using Mycobacterium tuberculosis Genomes
- 2025
- 2025
ICLR 2025 MLGenX Workshop Workshop paper
Beyond Sequence-Only Models: Leveraging Structural Constraints for Antibiotic Resistance Prediction in Sparse Genomic Datasets
Talks & media
Watch
Since 2025
Recent milestones
- NextPreparing to propose my dissertation, "Integrating Biological Signal to Improve Generalizability and Interpretability of Machine Learning in Genomics," in early Fall 2026.
- 2026Submitted the FARM manuscript (biophysics-aware resistance forecasting) — under review at PNAS.
- 2026Selected as a fully funded graduate participant for the Tapia Conference.
- 2026Featured as a graduate researcher in UMass's AI at UMass public-engagement campaign.
- 2025–26Serving as elected Ph.D. Graduate Representative, Faculty Senate / CICS, UMass Amherst.
- 2026Reviewing for Bioinformatics Advances and MLCSB (ISMB).
- 2025Published a research article in eLife on predicting antibiotic resistance using protein structural context.
- 2025Presented work at the ICLR MLGenX workshop and delivered a spotlight talk at MLCB.
- 2025Completed my Master's in Computer Science and advanced to Ph.D. candidacy.
Contact
Research & collaboration
Want to talk research, collaboration, or TB genomics? Book 30 minutes, no agenda required.