Work across our Kubernetes / GCP stack to deploy and scale model serving, with a focus on latency, reliability, and cost.

Implement features around billing, usage tracking, and metrics (PromQL, BigQuery) to give customers better observability and control.

Collaborate with frontend (TypeScript / React) and product to ship end‑to‑end improvements to the developer experience.

Designed and implemented features in a production codebase, from schema design to deployment.

Worked in a small team environment, owning tasks end‑to‑end and improving reliability and performance of existing systems.

Designed and trained segmentation models and data augmentation methods to handle heterogeneous clinical and low‑field MRI data.

Co‑authored IEEE ISBI and Nature Communications work on portable MRI for Alzheimer’s disease and WMH quantification; contributed open‑source code to the WMH‑SynthSeg tooling: https://ieeexplore.ieee.org/abstract/document/10635502

Integrated image quantification tools into a clinical workflow: PACS/DICOM ingestion, preprocessing, radiomics feature extraction, and ML models for computer‑aided prostate cancer assessment.

Evaluated automatic quantification and reporting to support hospital‑specific normative values and pathology detection.