PAIMOD.PAI-08 · v1.0

Embody AI.
Move things in the
real world.

8 micro-lessons · ~84 min · Real Docker images

WORLD-MODEL ENCODER
FRAME 0042 · POSE LOCKED
WORLD MODEL · TICK 0.04s
X
Y
Z
X-AXIS 23°
Y-AXIS 145°
Z-AXIS 0°
TELEMETRY · 25 Hz · sim-to-real
PAIAI ENGINEERINGHOT

Physical AI · robotics, world models & VLA

World models, VLA policies, and the on-robot runtime — the 2026 stack for embodied systems.

WHY THIS MATTERS · NVIDIA GTC MARCH 2026 · HUGGING FACE LEROBOT 0.5 · PHYSICAL INTELLIGENCE PI0.5
Physical AI was the centerpiece of NVIDIA GTC 2026 — Cosmos foundation world models (2M+ downloads), Isaac GR00T N1.7 (April 2026), Jetson Thor at $3,499. Hugging Face LeRobot 0.5 (March 2026) ships open Pi0.5 + OpenVLA-OFT + RTC. The 2026 stack — Isaac Lab 2.3 + Cosmos + GR00T + LeRobot v3 + ROS2 Jazzy — replaces the bespoke 2023 robot ML pipelines.
WHAT YOU'LL LEARN
01The three layers of Physical AI
02Sim-first development with Isaac Lab
03VLA policies — Pi0.5, OpenVLA, GR00T
04Data pipeline — LeRobotDataset v3 + teleop
05Sim-to-real — DR + sysID + DAgger
06ROS2 Jazzy + DDS deployment
07Safety, eval & guardrails
08World models, Cosmos & synthetic data
YOU'LL BE ABLE TO
Decompose any Physical AI system into 8 layers and debug at the right one
Train a contact-rich Isaac Lab 2.3 policy with PPO + DR + ONNX export in 30 minutes
Wrap a Pi0.5 / OpenVLA / GR00T policy in a ROS2 Jazzy + Cyclone DDS + Foxglove + safety-guardian runtime
SKILLS YOU'LL GAIN

Real skills, real career delta.

Skills you'll gain

10
  • Read VLA / world-model papers without panicWorking

    Decode Pi0.5, OpenVLA, GR00T N1.7, Cosmos Predict/Transfer/Reason, RT-2, Diffusion Policy, RDT — what the architecture is, what dataset, what eval suite, where the gap is.

  • Build & playback a LeRobotDataset v3Working

    Record episodes with `lerobot-record`, push to Hugging Face Hub, replay with `lerobot-replay`, segment by language. The 2026 data lingua franca.

  • Train a contact-rich manipulation policy in Isaac Lab 2.3Production

    PPO + domain randomization for a Lift / Open-Drawer / Stack task; vectorized sim on a single GPU, Hydra config, Wandb logging, ONNX export, Lab-Arena gating.

  • Deploy a VLA policy server (Pi0.5 / OpenVLA-OFT / GR00T N1.7)Production

    FastAPI + ONNX/TensorRT in Docker; sub-50ms p99 latency on RTX 4090 / Jetson Thor; horizon=50 action chunks; Pydantic observation contract.

  • Wire a learned policy into ROS2 JazzyProduction

    rclpy node, action server, JointState publisher, Foxglove 2.0 panel layout, Cyclone DDS QoS profile that doesn't drop frames at 50Hz.

  • Run the standard eval suites (LIBERO, SIMPLER, RoboCasa, Lab-Arena)Working

    Author the eval config, run on sim, parse the success-rate JSON, gate CI on regressions, report MMRV + Pearson against real rollouts.

  • Domain-randomize + sysID + DAgger closure for sim2realProduction

    DR ranges in Isaac Lab event-manager API; MuJoCo MJX system identification (mass/friction/PD gains); DAgger with real teleop expert; the standard 5-step handoff.

  • Wrap any policy in a safety guardianWorking

    Rate limiter on joint targets, joint-position/velocity clamps, force/torque thresholds, soft e-stop watchdog, ISO 13482-aligned event log; works in sim and real.

  • Generate synthetic episodes with NVIDIA CosmosAdvanced

    Cosmos Predict 2.5 rollouts; filter with Cosmos Reason 2; merge synth + real in LeRobot v3; quantify sim-only vs hybrid eval lift on RoboCasa or LIBERO.

  • Run a fully on-device runtime on Jetson Thor / OrinAdvanced

    TensorRT NVFP4/FP8 quantization for the VLA policy, CUDA-aware DDS, Foxglove over Wi-Fi telemetry, thermal/power budgeting — the deployment shape humanoid startups buy.

RUNNABLE ON YOUR MACHINE
$ docker pull snap/physical-ai:stack-tour
$ docker run --rm -it snap/physical-ai:stack-tour
snap/physical-ai:stack-tour
QUICK PREVIEW · 7 MIN
VERIFIED ENGINEER REVIEWS
The sim-to-real evaluator alone earned its keep — we now gate fleet rollouts on SIMPLER MMRV.
@robot_rinVERIFY ON GITHUB
Safety-guardian compose is exactly the audit trail our cert auditor wanted. Saved us 3 weeks.
@kofi.infraVERIFY ON TWITTER
LESSONS8
HOURS~1.4
LEARNERS720
THIS WEEK+49%