Chest X-Ray Triage Neural Network
Built a calibrated CNN pipeline that ranks urgent studies, reports uncertainty, and exports a clinician-readable model card.
About and Skills
I build Data Science systems in Python, train Neural Networks, evaluate recommender systems, and ship model monitoring dashboards for real users.
PythonNeural NetworksData SciencePyTorchFeature StoresMLOps
Projects
Real-Time Fraud Graph Model
Combined graph embeddings and gradient boosted features to flag suspicious payments with deterministic replay tests.
LLM Search Relevance Evaluator
Created a Python evaluation harness for retrieval quality, prompt regression tests, and offline Data Science review.
Original Twist: Failure Museum
Selected lesson: data leakage hides inside timestamp joins.
Current Focus
Monitoring drift for neural recommender models.