Deploying ML Models and the Things That Go Wrong
summary
Marianna Diachuk discusses practical pitfalls in deploying machine learning models, contrasting teams that have dedicated deployment engineers with those that expect data scientists to handle modeling, deployment, and monitoring. She shares lessons learned since 2017 and concrete anecdotes about what goes wrong during development and production deployment of ML systems.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 2:18Deployment infrastructure
- 5:28Production ML pitfalls
- 9:00Debugging model performance
- 12:30ML team collaboration
- 16:44Optimizing ML workflows
- 20:19Monitoring models
- 23:35Model lifecycle
- 26:48Lessons learned
- 28:58Q&A
related talks
42mDeploying ML models and all the things that go wrong
39mApplying deployment oriented mindset for building Machine Learning models
06mMachine learning is about your data and deployment
30mChallenges of machine learning: The new ML lifecycle
52mThe ML Lifecycle in Combat Conditions
04mManaging Machines
47mMLOps tools to automate ML deployment
46mI know Machine Learning but... am I ready for the real world?