Explore the intricacies of MLOps in this 50-minute podcast episode featuring Katrina Ni and Aaron Maurer, discussing the build or buy dilemma and comparing startup vs. enterprise approaches. Delve into Slack's machine learning journey, including their recommender system framework, data privacy challenges, and innovative solutions for cold start problems. Learn about the evolution of ML roles, the importance of diverse skill sets, and predictions for the future of machine learning. Gain insights on model serving, MLOps maturity levels, and the primary difficulties faced in implementing ML systems at scale.
MLOps Build or Buy - Startup vs. Enterprise Perspectives
Overview
Syllabus
[] Aaron and Katrina's preferred coffee
[] Recommender and System and Jake
[] Takeaways
[] Introduction to Aaron Maurer & Katrina Ni
[] Aaron Maurer & Katrina Ni's Recommend API blog post
[] 10-pole machine learning use case and Rex's use case
[] Genesis of Slack's recommender system framework
[] The Special Sauce
[] Speaking the same language
[] Use case sources
[] Slack's feature engineering
[] Main CTR models
[] Data privacy
[] Slack's recommendations problem
[] Fine-tuning the generative models
[] Cold start problem
[] Underrated
[] Baseline
[] Cold sore space
[] LLMs in Production Conference Part 2 announcement!
[] Data scientists transition to ML
[] Unicorns do exist!
[] Diversity of skill set
[] The future of ML
[] Model Serving
[] MLOps Maturity level
[] AWS Analogy
[] Primary difficulty
[] Wrap up
Taught by
MLOps.community
