Skip to content

Responsible AI Track

This track combines fairness diagnostics, explainability methods, and causal reasoning for safer model-driven decisions.

  1. projects/xai-fairness-audit-showcase
  2. projects/causalml-kaggle-showcase
  3. projects/sota-supervised-learning-showcase (evaluation extension)

Core Skills Covered

  • Explainability with SHAP/LIME.
  • Subgroup fairness analysis and mitigation tradeoffs.
  • Causal treatment effect estimation and policy simulation.
  • Interpretation of model decisions under uncertainty.

Evidence Artifacts To Inspect

  • explainability outputs in artifacts/explainability/
  • subgroup fairness reports in XAI/fairness artifacts
  • uplift and policy simulation outputs in causal showcase

Suggested Reflection Prompts

  • When is an explainable model still unsafe to deploy?
  • Which subgroup metric is most meaningful for this use case?
  • How does causal uplift change action policy compared with pure prediction?