AI & ML — EVALUATION, SAFETY & PRODUCTION

Evaluation, Safety & Production

Most topics here have three levels — Introduction, Intermediate, and Advanced — so pick one and go as deep as you want. A few go straight to Intermediate for readers who already have the basics down.

Evaluating LLM Outputs
Automated metrics, human evaluation, and LLM-as-judge — what each one measures, and the real biases to guard against.
Intermediate Advanced
AI Safety & Guardrails
Why one defense layer is never enough — input filtering, output filtering, system-prompt limits, jailbreaks, and how labs actually red-team before shipping.
Intermediate Advanced
Production Monitoring & Observability
What to log, how to catch drift before users do, and how to respond when an LLM system breaks in production.
Intermediate Advanced