fairlearn

Auditing algorithmic bias in criminal justice, hiring, lending, healthcare, welfare, and tenant screening: 7 open-source audits, measurable fairness gaps, and concrete fixes.

Personal Data Science Projects

Demos about Teaching your Models to play fair with FairLearn

Assess fairness of machine learning models and choose an appropriate fairness metric for your use case with Fairlearn

Learn different techniques for mitigating fairness-related harms using Fairlearn.

An end-to-end MLOps pipeline for a production-grade fraud detection model. This project demonstrates best practices including data versioning (DVC), experiment tracking (MLflow), CI/CD (GitHub Actions), containerization (Docker), deployment on GKE, and advanced model analysis (poisoning attacks, drift, fairness, explainability).

🥉 3rd Place Hackathon Winner | Solo-built AI platform for detecting, explaining, and mitigating bias in machine learning systems.

Drop-in encrypted Fairlearn metrics over CKKS. Same API surface; ciphertext arithmetic via TenSEAL or OpenFHE.

Microsoft Ignite - Getting started on your health-tech journey using responsible AI

🔬 Drop in any ML model → get SHAP explainability, fairness audit & drift detection in seconds

Demo's of FairLearn and InterpretML as described in my article on responsible AI.

AI-powered bias detection for datasets and ML models — with fairness metrics, natural language reports, and explainability tools.

Project: Credit XAI – Ethical AI in Lending & Red Teaming AI Credit Underwriting

Fairlearn Jupyter notebooks samples

An ethically-aware deep learning project to predict credit card offer acceptance while mitigating income-based bias using SHAP, Fairlearn, and AIF360.

AI-powered bias auditing tool that detects demographic disparities in datasets and explains findings in plain English using Gemini.

A platform developed with Cash App to help ML engineers detect and visualize biases in models using Fairlearn. Features include a collaborative and interactive dashboard (React, Chart.js), a Flask backend, and a secure MySQL database for data storage and analysis.

This repository was used for my thesis. The goal was to find a biased dataset, and mitigate its bias. That is done under the patients directory. Check the README file for more.

End-to-end MLOps pipeline for heart disease classification, deployed on GKE with CI/CD, explainability, fairness testing, and drift detection.

Replication of bias mitigation experiments from arXiv:2412.09896 - Fairlearn vs AIF360 on the Adult Income dataset with XGBoost