Agentic AI Curriculum
A linear, job-ready path — modules of short lessons, interactive code, and checks.
Foundations
The agent loop, the LLM as decision policy, and the deterministic mock-LLM harness every later module builds on.
Prompting & Control
Turning model output into reliable, parseable decisions.
Tool Use
Letting agents act through well-described, testable tools.
Retrieval & RAG
Grounding answers in retrieved context with citations.
Memory & State
Carrying state across turns: working memory and summaries.
Planning & Reasoning
Decomposition, plan-execute, and reflection loops.
Multi-Agent Systems
Coordinating specialised agents: supervisor and workers.
Evaluation & Testing
Measuring agents: task success and trajectory checks.
Guardrails & Safety
Validating inputs/outputs and executing tools safely.
Cost, Latency & Reliability
Budgets, caching, timeouts, retries, and fallback.
Observability & Debugging
Tracing the loop and replaying failed trajectories.
Production & Deployment
Statelessness, real-API integration, rollout, capstone.