Training Catalogue
Our Training Programmes
Seven intensive programmes covering every layer of the modern robotics stack — from ROS 2 fundamentals to cutting-edge Language Action Models. All delivered on dedicated GPU cloud instances.
ROS 2 Fundamentals
ROS 2 Fundamentals is the definitive starting point for engineers new to the Robot Operating System 2. You will learn the complete ROS 2 architecture — from workspace setup to deploying a simulated mobile robot in Gazebo. This course is 100% hands-on: every lecture is paired with a lab exercise running on a dedicated GPU cloud instance.
- ✓ Set up a full ROS 2 Humble workspace with colcon build tools
- ✓ Implement nodes, topics, services, and action servers from scratch
- ✓ Simulate a differential-drive robot in Gazebo Classic and Gazebo Sim
NVIDIA Isaac Sim Essentials
NVIDIA Isaac Sim Essentials gives you a thorough grounding in the world's most powerful robot simulator. Powered by PhysX and Omniverse, Isaac Sim enables photorealistic rendering, physically accurate simulation, and seamless ROS 2 integration — all on GPU. You will leave with the skills to build, sensor-equip, and run full robot simulations from scratch.
- ✓ Install and configure Isaac Sim on a cloud GPU instance
- ✓ Import URDF and USD robot models into Isaac Sim
- ✓ Configure LiDAR, RGB-D camera, and IMU sensor suites
Digital Twin & Simulation
Digital Twin & Simulation teaches you to build live virtual replicas of physical robot systems using NVIDIA Omniverse and USD. You will learn how to calibrate physics to match real hardware, stream sensor data between the physical and virtual world, and integrate digital twins into CI/CD pipelines for continuous testing. This course is ideal for engineers who already understand simulation basics and are ready to go production-grade.
- ✓ Architect a complete digital twin for a real robot cell
- ✓ Import CAD assets and configure physics materials accurately
- ✓ Calibrate joint friction, damping, and inertia from real data
Physical AI & Robot Learning
Physical AI & Robot Learning bridges the gap between deep reinforcement learning theory and real robot deployment. Using NVIDIA Isaac Lab (built on Isaac Sim), you will design Gym-compatible environments, train locomotion and manipulation policies with state-of-the-art RL algorithms, and apply sim-to-real transfer techniques so learned policies work on physical hardware. This is the course for engineers ready to push robots beyond pre-programmed routines.
- ✓ Design custom Gym environments inside Isaac Lab
- ✓ Implement PPO, SAC, and TD3 for continuous control tasks
- ✓ Vectorise environments across thousands of parallel simulation instances
NVIDIA Omniverse & USD
NVIDIA Omniverse & USD dives deep into the platform that underpins Isaac Sim and the broader robotics simulation ecosystem. You will learn Universal Scene Description (USD) authoring, Omniverse Nucleus for collaborative workflows, custom Kit extension development, and Omniverse Connectors for integrating with CAD tools like SolidWorks and Blender. The course also covers RTX rendering for photorealistic synthetic data generation.
- ✓ Understand Omniverse architecture: Kit, Nucleus, RTX Renderer
- ✓ Author and compose USD scenes programmatically with Python API
- ✓ Build custom Omniverse Kit extensions in Python
Language Action Models (LAM)
Language Action Models (LAM) represents the frontier of physical AI — systems that understand natural language instructions and translate them into physical robot actions. In this advanced course you will study the architecture of foundation models for robotics (including NVIDIA GR00T), implement instruction-following pipelines, and deploy inference workloads on Isaac Sim. You will work with the very techniques shipping in next-generation autonomous robots.
- ✓ Understand transformer architectures adapted for robot control
- ✓ Study GR00T, RT-2, and OpenVLA foundation model designs
- ✓ Implement instruction-conditioned policy inference pipelines
NVIDIA GPU Stack & CUDA
NVIDIA GPU Stack & CUDA takes you under the hood of the compute layer powering every NVIDIA robotics product. You will write CUDA kernels, optimise neural network inference with TensorRT, serve models at scale with Triton Inference Server, build video analytics pipelines with DeepStream, and master GPU memory management and profiling tools. This course is for engineers who need to squeeze maximum performance out of NVIDIA hardware.
- ✓ Write and debug custom CUDA C++ kernels for robot data processing
- ✓ Optimise neural networks with TensorRT: FP16 and INT8 quantisation
- ✓ Deploy model serving at scale with Triton Inference Server
Need Custom Enterprise Training?
Book multiple seats across programmes, get a dedicated instructor, and customise lab scenarios to match your specific robot hardware and use cases. Volume pricing available.
Request Enterprise Package →