I study data‑efficient reinforcement learning, offline/continual learning under non‑stationarity, and domain adaptation. I analyze dataset effects, develop credit‑assignment and curriculum/distillation methods, and build interpretable, logic‑guided models and LLM tooling.