My work sits at the intersection of computational genomics and immunology. I design methods that turn raw sequencing data into decisions: which viral integration events matter, which mutated peptides an immune system can actually see, and how to read a genome fast enough to matter in the clinic.
On the systems side, I built GenoCache, a learned-embedding alignment engine that reaches 95 percent locus accuracy at 9 ms per long read on an A100, within 0.85x of minimap2's three decades of hand-tuned heuristics. On the biology side, I built an end-to-end neoepitope discovery pipeline for HPV-driven cancers, and I am extending it across TCGA cohorts to ask whether integration burden and host gene expression predict survival.
None of it is only computational. I have run the assays, sequenced the samples, and validated in the field, and I spent two years developing vaccine potency assays under cGMP at Merck. I care about methods that survive contact with real, messy biological data.
PyTorchCUDAFAISSHyenaDNA / DNABERT-2scGPTpVACtools / NetMHCpanVIRUSBreakend / GRIDSSOxford NanoporeScanpy / SeuratPython & R
FocusCancer immunogenomics & ML for genomics
Ph.D.Biological Design, ASU (2025)
NowPostdoc, Biodesign Institute
IndustryScientist II, Merck (cGMP)
Based inTempe, Arizona
Patents2 filed · 2 pending
Publications5 peer-reviewed · 5 in review/prep