Smart Range Hood Controller: ML Onset Detection & Physics-Based MPC Optimization
AI/IoT platform for predictive indoor air-quality management and energy efficiency, combining machine learning, physics-based simulation, optimization, and closed-loop control.
Key Technical Highlights:
__________________________
• Proactive Control & AI — Architected a hybrid IoT platform combining physics-informed ML, Model Predictive Control (MPC), and Pydantic-validated LLM scenario generation to optimize indoor air quality and energy consumption.
• MLOps & Experimentation — Built reproducible DVC data pipelines, MLflow experiment tracking, and Optuna hyperparameter tuning, managed via uv.
• Production API & UI — Engineered a FastAPI REST API layer orchestrating core simulation services with an interactive Streamlit frontend.
• Testing & CI/CD — Automated multi-tier Pytest suites (unit & integration) via GitHub Actions, publishing multi-service Docker images to GHCR.
