Agent coding
Build the pieces of an LLM agent in Python — the agent loop, tools, retries, retrieval, and evals — the coding round for AI-native roles. The model calls are mocked, so the exercise is the control logic. 23 problems across 6 patterns.
The Agent Loop
0 / 4 solved
Tools & Function Calling
0 / 4 solved
Reliability & Retries
0 / 4 solved
Retrieval (RAG)
0 / 4 solved
Memory & Context
0 / 3 solved
Evaluation
0 / 4 solved
Show the dependency map
✓ Track complete — every pattern solved. Nicely done.
Follow the arrows: finish a topic to unlock what builds on it. Progress is saved on this device.