Fairlytics

• 

  High level architecture

• 

  Counterfactual testing to see if a model performs the same while flipping the feature values

• 

  Interpretability of a model to check if there is a feature being focused on overtly

• 

  Landing page

Inspiration

No real time bias monitoring for AI predictions

What it does

Real time fairness monitoring to mitigate bias

How we built it

docker, flask, sklearn, grafana

Challenges we ran into

integrating multiple services

Accomplishments that we're proud of

were able to build everything from scratch with a small team

What we learned

What's next for Fairlytics

scale to multiple models, integrate other features like user feedback loop

Built With

• docker
• flask
• grafana
• postgresql
• python
• sklearn