Build robotics AI systems
Action playbook from six AI Engineer / World's Fair talks. Do not summarize talks — pick a workflow and execute it.
Supporting files (read when needed):
- workflows.md — workflows A–F (steps, deliverables, stop conditions)
- source-index.md — src-NNN → talk learnings in ingest-into-skills
Optional: {SKILL_OUTPUT_DIR}/build-robotics-ai/
Step 0 — Pick workflow
Use the decision tree below. Open the matching section in workflows.md.
What is the user trying to do?
├─ Closed-loop lab / co-scientist agent → A
├─ Humanoid systems performance (latency/data) → B
├─ Commercial kitchen/service robotics → C
├─ AV + VLM planning (Waymo EMMA) → D
├─ Adopt humanoid foundation model (GR00T) → E
└─ Ship affordable open humanoid platform → F
Stop summarizing once a workflow is identified — run its checklist.
Install
cp -r skills/build-robotics-ai ~/.claude/skills/
cp -r skills/build-robotics-ai ~/.cursor/skills/
cp -r skills/build-robotics-ai ~/.codex/skills/
From skills-i-use or ingest-into-skills (playlists/robotics-autonomy-ai-engineer/).
Cross-cutting rules
| Rule | Source |
|---|---|
| Profile end-to-end control latency | [src-002 @ 7:11] |
| Sim/real eval before planner changes | [src-004 @ 14:16] |
| Teleop safety before full autonomy | [src-006 @ 10:29] |
Disputed steps: see source-index.md. Name workflow A–F; save artifacts to ./skill-outputs/build-robotics-ai/ when requested; do not auto-commit.
Invocation examples
@build-robotics-ai humanoid data pipeline architecture
lab agent with live sensors