Guided learning path
Agent Performance
Understand tool-call execution, cache behaviour, retries, and trajectory evaluation in stateful AI agents.
Who this path is for
Engineers building agents that call tools, retry operations, or maintain long-running context.
What you will learn
Trace the work around a tool call; identify recomputation and latency costs; handle ambiguous tool outcomes; evaluate the execution trajectory.
Prerequisites
Familiarity with model APIs, tool calling, and basic request/response services. Read the model-internals cache material if KV-cache concepts are new.
Difficulty and workload
Intermediate to advanced. Trace one agent execution by hand before using the linked articles to examine performance and reliability boundaries.
Suggested reading sequence
- When an Agent Calls a Tool, What Happens to Its KV Cache?
- Make Agent Tool Retries Safe With Typed Contracts and Idempotency
- Evaluate the Agent Trajectory, Not Just Its Final Answer
This reading sequence uses existing articles. New guided lessons will appear in the published-lessons list below. Return to all learning paths to explore a related subject.
Curriculum
Published lessons
- No lessons have been published in this path yet.