Clinical AI Builder
I build interpretable, reliable AI that clinicians trust.
From raw EHR to GPU‑accelerated pipelines and bedside‑safe decisions, built with first‑principles thinking from physics → ML/DL → systems.
Learn more
Principles
- • Safety, explainability, and validation > model vanity.
- • Privacy‑by‑design and least‑privilege data access.
- • Human‑in‑the‑loop deployment; clinicians at the center.
Skills
- • ML/DL: CV · NLP · RL
- • Frameworks: PyTorch/TensorFlow, scikit‑learn, NumPy
- • Data & Scale: SQL, Spark; batch/stream pipelines
- • Systems: containers (Docker), orchestration, CI/CD
- • Acceleration: CUDA & multi‑GPU training
- • Healthcare data: EHR (tabular/text), clinical notes, MIMIC‑III
- • MLOps: experiment tracking, model monitoring & drift handling
