AI curriculum

Course outline for AI Observability

AI failures can arise from prompts, models, or tools. Trace an application interaction to connect resource usage, execution steps, and decision evidence without exposing sensitive inputs.

Technology: AI Operations, Evaluation and Observability
Duration: 6 hours
Modules: 3 · 2 hours each
Level: Prerequisites listed

About AI Observability

AI failures can arise from prompts, models, or tools. Trace an application interaction to connect resource usage, execution steps, and decision evidence without exposing sensitive inputs.

AI Observability Course Objectives

Define trace fields for model and tool interactions.
Analyze token use, latency, and cost assumptions.
Reconstruct a failed agent decision from recorded evidence.

Pre-requisites

Familiarity with LLM applications and agent tools.
Basic logging or observability experience.

Lab Setup

Computer with a trace viewer or analysis notebook.
Use synthetic prompts and supplied traces or local mock endpoints; no production telemetry access is needed.

Detailed Course Outline

Proposed modules

Module 1: Model interaction traces

Prompt traces
Model traces
Token usage

Practical outcome: Record an interaction with safe metadata and input redaction.

Module 2: Resource and failure analysis

Latency
Cost
Failure analysis

Practical outcome: Identify a costly or slow execution stage from sample traces.

Module 3: Agent execution evidence

Agent traces
Tool calls
Decision provenance

Practical outcome: Reconstruct which evidence and tools influenced a decision.

Practical exercise

Analyze a failed agent trace and propose instrumentation and redaction improvements.

How we train

Ask about exercises, instructor feedback and the prior knowledge you need. Tool-specific courses marked provisional may become modules in a broader course.

Discuss your learning goals