The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Machine Learning at GitHub with Omoju Miller - #313
8 snips
Oct 31, 2019 
Omoju Miller, a Senior Machine Learning Engineer at GitHub, shares her insights on enhancing developer collaboration and open-source participation. She discusses her dissertation on computer science education and her pioneering work on GitHub's machine learning team. The conversation dives into the trends of GitHub stars for ML packages, reflecting on the shifts in the industry. Omoju also highlights the challenges of integrating new technologies and the exciting potential of open-source innovation, inspiring collaboration and creativity in the coding community.
AI Snips
Chapters
Transcript
Episode notes
Omoju's ML Journey
- Omoju Miller's machine learning journey began in the late 90s, working with C++ and MATLAB.
- She transitioned to knowledge representation and semantic web projects before returning to ML when robust ecosystems emerged.
Joining GitHub's ML Team
- Omoju Miller joined GitHub's newly formed machine learning team.
- Her goal was to leverage GitHub's vast data to enhance developer collaboration and reduce transaction costs.
GitHub's ML Approach
- GitHub uses data to enhance features like repository labels and recommendations.
- They utilize abstract syntax trees (ASTs) and other techniques for code representation and analysis.