Reinforcement Learning & Control

• Policy gradients, value-based RL, world models
• Planning and model‑based control (MPC)
• Isaac Gym / MuJoCo environments

Perception & Multimodal

• Vision + language + proprioception fusion
• Visual servoing, depth fusion, tactile sensing

Sim2Real & Embodiment

• Domain randomization & transfer
• Calibration and real‑world evaluation

Research Practice

• Reproducing landmark papers (DreamerV3, Gato, Gemini)
• Clean repos with baselines and ablations