CV / Background

Doctoral research aimed at programmable cancer immunotherapies and spatial immune modeling.

Integrated data science frameworks with advanced molecular biology to prototype translational pipelines.

Built the computational tooling foundation that powers Kelvin's current lab work.

• Characterized CAR T exhaustion and myeloid biology for the first on-treatment scRNA-seq cohort in LBCL, surfacing biomarkers for new trials.
• Predicted CD19-CAR response and toxicity using multimodal ML classifiers on pre-infusion immune features, integrating sgGPT-based transformers.
• Mapped temporal and spatial programs in GD2-CAR therapy for diffuse midline glioma using CosMx, CITE-seq, and proteomic readouts.
• Built the lab's single source of truth for patient metadata, wet-lab instrumentation, and ML pipelines, enabling real-time translational decisions.

• Implemented Raman spectroscopy-based feedback control for cell culture bioreactors, enhancing feed control capabilities.
• Delivered dashboards and automation that reduced manual interventions during pilot biologics runs by 80%.

• Built ranking models on structure-aware embeddings to triage antibody screening campaigns.
• Deployed decision-support tooling that accelerated hit triage for wet-lab partners.

• Built accelerator-aware inference pipelines for large-scale biological simulations and HPC clusters.
• Published internal whitepapers to senior leadership on GPU/CPU utilization for omics workloads.

• Constructed tension-area hysteresis analyses and impedance-based manometry tooling for EGJOO patients.
• Led data ingestion for clinical swallow studies and mentored undergraduates in reproducible analytics.

• Shipped early prototypes for cfDNA analytics and assay QC dashboards.
• Anchored Kelvin's transition from pure wet-lab work into computational biology.