research Motor Thermal Modeling with Deep Learning Recurrent (LSTM, GRU) and feed-forward virtual sensors for online motor temperature estimation on the public Paderborn PMSM dataset. Methodology generalizes my industrial work on rare-earth-free EESM motors to an open benchmark. Hybrid SAC + MPC for Vehicle Lateral Control A closed-form blend of a Soft Actor-Critic policy and a constrained linear MPC for connected-and-automated-vehicle lateral control. Combines the tracking quality of deep RL with the actuator envelope and interpretability of MPC. tools doc-agent — Simulink Documentation An LLM agent that reads Simulink models and writes their documentation — grounded in tool calls against the real model, exported to HTML, DOCX, and PDF.
