Cybera Fellowship


  Project Description
  Blog Post

Technologies

  • Docker
  • Agile Frameworks
  • Numpy / Pandas / Scikit-Learn / Jupyter
  • Json API scrapping.

Summary

The goal of this project was to build a recommender system to optimize career paths for students athletes interested in pursuing a professional career in hockey. I build a Markov Decission Process (MDP) using historical data from the EliteProspects API to calculate the transition probability between hockey leagues. This model works as a recommender system that helps student athletes to choose optimal career paths in hockey based on athletic training, academic paths, and financial measurables.

What did I do?

  • Developed an automated data extraction library for capturing, cleaning, and archiving 250,000 records from upstream API.
  • Researched and built a reinforcement learning ML model from scratch using Markov Decision Processes.
  • Provided data visualization solutions to enable communication between technical team and industry partners.
Back to Top ↑