Skip to content

👨‍💻 Developer Path

L100 L200 L300 L400

You write Python or C# and want to build AI agents with full code control — from a simple chatbot to a production multi-agent system.

Your Journey

L100: Learn the basics          L200: Build real agents         L300: Production patterns        L400: Expert
GitHub Models (free)      →     MCP servers, RAG, Agent    →    Observability, security,    →    Multi-agent, capstone
Foundry Local (free)            Framework, guardrails           deploy, evaluation               architecture

Phase 1 — Foundations (free, no Azure needed)

Order Lab Title Level Time
1 Lab 004 How LLMs Work L50 ~20 min
2 Lab 013 GitHub Models — Free LLMs L100 ~25 min
3 Lab 078 Foundry Local — Run Models Offline L100 ~45 min
4 Lab 017 Structured Output & JSON Mode L100 ~25 min
5 Lab 018 Function Calling & Tool Use L100 ~30 min

Phase 2 — Build Real Agents

Order Lab Title Level Time
6 Lab 076 Microsoft Agent Framework L200 ~75 min
7 Lab 020 Build an MCP Server in Python L200 ~45 min
8 Lab 022 RAG with GitHub Models L200 ~50 min
9 Lab 072 Structured Outputs Reliability Benchmark L100 ~45 min
10 Lab 082 Agent Guardrails L300 ~75 min

Phase 3 — Production Patterns

Order Lab Title Level Time
11 Lab 049 Agent Tracing with OpenTelemetry L300 ~75 min
12 Lab 037 CI/CD for AI Agents L300 ~45 min
13 Lab 036 Prompt Injection Defense L300 ~40 min
14 Lab 038 AI Cost Optimization L300 ~45 min

Phase 4 — Expert

Order Lab Title Level Time
15 Lab 074 Foundry Agent Service L300 ~120 min
16 Lab 055 A2A + MCP Capstone L400 ~120 min
17 Lab 084 Capstone — OutdoorGear Agent L400 ~180 min

Total: 17 labs · ~17 hours · From zero to production-ready agent builder

Phase 2 additions

Use Labs 090, 092, 093, 096 and Challenge 010 for tested, secure, agentic delivery.

Phase 3 route

Use Labs 102–106, the modernized Agent Framework/MCP/Foundry labs, and the Instructor Hub delivery formats appropriate to this role.