Projects
Research implementations, production ML systems, and technical explorations
Collaborative Nested Learning
Extended Google's NeurIPS 2025 research with 5 timescales and 9 bidirectional knowledge bridges. Novel approach to continual learning that prevents catastrophic forgetting.
Clinical RAG Platform
Production retrieval-augmented generation for healthcare decision support. Multi-tenant system serving 180+ hospitals with safety-critical design.
Document Understanding
Enterprise HIPAA-compliant document extraction with multi-provider LLM integration (Claude, GPT-4, Azure OpenAI), human-in-the-loop workflows, and reinforcement learning from human feedback.
Mellifera: Agricultural ML Platform
Production ML classifiers for precision beekeeping: treatment timing (92% accuracy), winter survival prediction (0.87 AUC), and feeding optimization. TensorFlow.js models under 25KB total, with plans to move to edge inference.
Past Explorations
Projects that shaped my technical direction
Technical Focus Areas
Continual Learning
Implementing frontier research on learning without forgetting. Multi-timescale optimization and knowledge preservation.
Production RAG
Safety-critical retrieval-augmented generation for healthcare. Hybrid retrieval, confidence scoring, citation tracking.
Multi-Agent Systems
Orchestrating autonomous agents in production. State management, failure handling, observability.
ML Infrastructure
Platformizing AI deployment with IaC patterns. Prompt registries, eval pipelines, versioning.