06. Causal Inference for AI Systems
AI Systems
Causal Inference
LLMOps
Lecture Notes
Causal methods for evaluating, deploying, monitoring, and governing AI products and human-AI workflows.
This course flips the direction: AI systems become the intervention. It covers estimands, experiments, triggered exposure, RAG and agent evaluation, routing bias, monitoring, fairness, cost-benefit, governance, and lifecycle management.
Notebook links open rendered HTML pages generated from the source notebooks under notebooks/lectures/. Code is visible by default; rendering is configured not to execute live notebook code, so local LLM or GPU-heavy cells are not triggered during website builds.
Notebook Sequence
- 01. AI Systems as Interventions
- 02. From Model Metrics to Causal Impact
- 03. Estimands for AI Product Changes
- 04. Randomized Experiments for AI Features
- 05. Triggered Experiments and Exposure Logging
- 06. Human-AI Workflows and Post-Treatment Variables
- 07. Guardrail Metrics for AI Systems
- 08. LLM-as-Judge and Measurement Bias
- 09. Causal Evaluation of RAG Systems
- 10. Causal Evaluation of AI Agents
- 11. Policy Changes, Routing, and Selection Bias
- 12. Observational Evaluation of AI Deployments
- 13. Difference-in-Differences for AI Rollouts
- 14. Synthetic Control for Enterprise AI Rollouts
- 15. Off-Policy Evaluation for Logged AI Decisions
- 16. Interference and Feedback Loops in AI Systems
- 17. Fairness, Bias, and Heterogeneous AI Effects
- 18. Causal Monitoring, Drift, and Regression Testing
- 19. Cost-Benefit and Decision Analysis for AI Deployments
- 20. Experiment Readouts for AI Product Teams
- 21. Governance, Auditability, and Model Lifecycle
- 22. Capstone: Evaluating an AI System End to End
How To Read This Track
- Work through the notebooks in order if you want the full course arc.
- Treat each notebook as a lecture plus lab: read the discussion, inspect the code, and rerun locally when you want to experiment.
- For AI-heavy notebooks, expect some brittleness when live model calls are enabled; that instability is part of the course material rather than something hidden from the reader.
The .ipynb sources remain in the matching folder under notebooks/lectures/.