I redesign lead candidates using both sequence- and structure-based protein design, and I’m building the pipeline behind a lab-in-the-loop design-build-test-learn cycle, so every round of bench data quickly feeds back into the next round of designs.

protein design structure-based design DBTL PyTorch HPC

Took ownership of the in silico side of antibody discovery and characterization, running the ML pipelines I’d built in earlier roles as production tools across therapeutic programs.

diffusion models structure-function modeling antibody engineering PyTorch HPC

I build diffusion-based pipelines for de novo antibody sequence design and custom ML tools for predicting affinity, stability, and manufacturability from proprietary wet lab data. The work is about turning model outputs into testable hypotheses for the bench team.

diffusion models structure-function modeling antibody engineering PyTorch HPC

Built a deep learning humanization platform for optimizing mouse-derived lead candidates and led the onboarding of Oxford Nanopore sequencing infrastructure, including the downstream analysis workflows.

antibody humanization deep learning PLM NGS pipelines

Built data pipelines from production databases to identify manufacturing bottlenecks and reduce turnaround times for gene therapy products. Got a thorough look at GMP environments and the quality standards that govern therapeutics.

data pipelines gene therapy GMP manufacturing

Transcriptomic analysis on Nanopore RNA-seq data and image analysis pipelines in ImageJ for characterizing microglial morphology. Contributed to work published in Cell Reports Medicine.

RNA-seq analysis Nanopore Sequencing cloud compute ImageJ neuroimmunology

Designed zebrafish behavioral enrichment experiments, performed gene enrichment analysis in R, and presented at multiple academic conferences.

behavioral neuroscience R gene enrichment zebrafish anatomy