Research, industry, and applied engineering roles.
Built backend microservices for an LLM-powered error-analysis product, improving reliability for high-volume operator and inference workflows.
Built a Django feature-management interface integrating LaunchDarkly across development, staging, and production environments, then containerized it for deployment in Kubernetes.
Built centralized cybersecurity risk-intelligence infrastructure across three relational data layers, combining AI-assisted evidence extraction, Bayesian ranking, attack-path generation, and an interactive React interface.
Built OCR-based real-time and batch document-ingestion and tax-form extraction services that converted multi-form uploads into validated structured JSON.
Teacher–student GANs for liver MRI synthesis and segmentation enhancement (6k+ slices).
Developed multimodal latent representations across 6,000+ patient records, comparing feature representations and 2D versus 3D embedding strategies to improve cohort separation and interpretability.
Compared LightGBM, LSTM, and Transformer models across 1M+ highly imbalanced credit-card transactions, varying feature representations to maximize fraud recall under a strict false-positive constraint.
Designed and delivered a multi-day applied-machine-learning curriculum for 100+ students through hands-on Python, real-world dataset, and model-building workshops.