SE4ML - Software Engineering for Machine Learning - Nadia Nahar

DataTalks.Club

We talked about:

• Nadia’s background
• Academic research in software engineering
• Design patterns
• Software engineering for ML systems
• Problems that people in industry have with software engineering and ML
• Communication issues and setting requirements
• Artifact research in open source products
• Product vs model
• Nadia’s open source product dataset
• Failure points in machine learning projects
• Finding solutions to issues using Nadia’s dataset and experience
• The problem of siloing data scientists and other structure issues
• The importance of documentation and checklists
• Responsible AI
• How data scientists and software engineers can work in an Agile way

Links:

• Model Card: https://arxiv.org/abs/1810.03993
• Datasheets: https://arxiv.org/abs/1803.09010
• Factsheets: https://arxiv.org/abs/1808.07261
• Research Paper: https://www.cs.cmu.edu/~ckaestne/pdf/icse22_seai.pdf
• Arxiv version: https://arxiv.org/pdf/2110.

Free data engineering course: https://github.com/DataTalksClub/data-engineering-zoomcamp

Join DataTalks.Club: https://datatalks.club/slack.html

Our events: https://datatalks.club/events.html