AI & LLM Agents Advanced • 10.5 hours total
Autonomous AI Agents & MCP Engineering in TypeScript
Go beyond simple prompt wrappers. In Hour 1 (free right here), we build a streaming tool-calling agent from scratch. In the remaining 9.5 hours, you will engineer custom MCP servers, sandboxed code execution, vector memory, and human-in-the-loop approval flows.
PRO
Sandboxed Execution, Human-in-the-Loop Approvals & Eval Suites
Unlock Full Course (10.5 hours total)
Unlock Full Course — $59 USDUnlock Full Course
10.5 hours total • 14-Day Money-Back Guarantee
⚡ Unlock Full Course
10.5 hours total • Cloudflare R2 HD Streaming • Source Code ZIP
Module 3: Production Evaluation & Guardrails
#4. Sandboxed Execution, Human-in-the-Loop Approvals & Eval Suites 📝 Lesson Architecture Notes & Code Branch
Git Branch: 04-agent-evalsAdds permission gates before mutating tool calls, automated trajectory grading, and latency/cost telemetry.
git checkout 04-agent-evals Local Git Branch