| Advanced Agents: Testing, Patterns and Production |
Failure taxonomy, trajectory evaluation and the testing layers from tools to multi-agent handoffs; LLM-as-judge and its calibration; logging, tracing and alerting; debate, hierarchical agents, recursive language models and vectorless RAG; evaluating deep-research agents; production architectures and scaling lessons |
Coming soon |
| Lab 1: Testing Agents |
Test a tool-using agent for a fictional company: trajectories and golden tests, tool tests, an LLM judge and adversarial cases; add structured logs, a trace waterfall and alerts; add a critic; then test the routing and handoffs of a multi-agent version and fix the bug the original tests missed. |
Coming soon |
| Lab 2 (Capstone): Build Your Own Hermes |
Build a small personal assistant modeled on Nous Research's Hermes Agent, one part at a time: an agent loop with tools, a chat gateway, long-term memory, skills the agent writes itself, scheduled jobs and an orchestrator with workers, then put it behind a live chat interface. |
Coming soon |
