Courses

Courses are the structured learning paths into robot learning and embodied AI — university lecture series with public materials, foundational online courses, and vendor training that walks through specific simulators or stacks. They cover the prerequisites (RL, control, perception) and the integration knowledge that papers alone rarely teach.

From an engineering standpoint, course materials are how you compress months of self-directed reading into a coherent mental model. They matter most for the parts of the field that are not in any single paper: the conventions, failure modes, and design trade-offs that experienced practitioners take for granted. For team onboarding and cross-discipline ramp-up, a curated course is almost always cheaper than ad-hoc reading.

When choosing, match the prerequisites you actually have (deep learning, control, RL) to the course's assumed background, and prefer courses with public lecture videos, problem sets, and reference code over slide-only materials. Pair one breadth course (robot learning overview) with one depth course (manipulation, RL, or locomotion) for the fastest practical ramp.

Start here

Spinning Up in Deep RL (OpenAI) is the most pragmatic on-ramp: clean explanations, reference implementations, and the right level of mathematical detail for engineers entering the field.

• CS 336 — Robot Learning (Stanford) — Modern robot learning, covering imitation and reinforcement learning.
• CS 224R — Deep RL for Robotics (Stanford) — Deep reinforcement learning for real-world robots.
• CS 287 — Advanced Robotics (Berkeley) — Planning, learning, and control for robotics.
• 16-831 — Introduction to Robot Learning (CMU) — CMU's foundations of robot learning.
• MIT 6.4210 — Robotic Manipulation — Perception, planning, and control for manipulation by Russ Tedrake.
• Spinning Up in Deep RL (OpenAI) — Educational resource covering deep RL fundamentals with reference implementations.
• Hugging Face Deep RL Course — Free hands-on course on deep reinforcement learning.
• NVIDIA DLI Robotics — Self-paced courses on Isaac Sim, ROS, and robot learning.
• CS 285 — Deep Reinforcement Learning (Berkeley) — Strong practical RL course with modern policy-gradient and model-based methods.
• Deep RL Bootcamp — Lecture series covering core deep-RL concepts and implementation practice.
• DeepMind x UCL RL Lecture Series — Advanced RL lecture track from leading research practitioners.
• 16-745 — Optimal Control and Reinforcement Learning (CMU) — Control and RL foundations for robotics deployment.
• MIT Underactuated Robotics — Open textbook/course on dynamics, planning, and control for underactuated systems.
• Fast.ai Practical Deep Learning — Applied deep-learning curriculum useful for perception and representation foundations.
• CS 234 — Reinforcement Learning (Stanford) — Core RL theory and algorithms with strong academic grounding.