Vincent MaladiereProba MLProba MLApply Conf 2022Apply Conf 2022Contributing to scikit-learnContributing to scikit-learnOpenCV TutorialOpenCV TutorialAboutAboutContactContact
  • Home
  • Day 1 - 🌅 Morning
  • 1. Managing the flywheel
  • 2. Lakehouse a new kind of platform
  • 3. ML for online prediction
  • Day 1 - ⚡ Lightning talks 1
  • 4. Why is ML hard
  • 5. DIY minimal feature store
  • 6. Enabling rapid model deployment in healthcare
  • 7. Extending Open Source feature Store
  • 8. Compass Composable & Scalable Signal Engineering
  • 9. Streaming is an implementation detail
  • 10. Effective system ML development
  • Day 1 - 🕶️ Afternon
  • 11. How to draw an owl and build effective ML stack
  • 12. Panel What engineers should know when building
  • 13 Is production RL at a tipping point
  • Day 1 - ⚡ Lightning talks 2
  • 14. Declarative ML Systems and Ludwig
  • 15. Accelerating model deployment velocity
  • 16. Semantic Layers & Feature stores
  • 17. Engineering for applied ML
  • 18. PyTorch’s next generation of data tooling
  • Day 2 - 🌅 Morning
  • 19. Faire’s journey toward modern data and ML stack
  • 20. Cash App’s real-time ranking ML system
  • 21. Feature engineering at scale with Dagger and Feast
  • Day 2 - ⚡ Lightning talks 1
  • 22. Data observability for ML team
  • 23. ML meet SQL
  • 24. Learning from monitoring more than 30 ML use-cases
  • 25. Lessons learned from working on Feast
  • 26. Evaluating RecSys in production
  • Day 2 - 🕶️ Afternoon
  • 27. Fire chat Is ML a subset or superset of programming
  • 28. Panel Alexander Ratner & Aparna Dhinakaran
  • 29. Intelligent Customer Preference Engine
  • Day 2 - ⚡ Lightning talks 2
  • 30. Are transformers becoming the most impactful
  • 31. Training large scale recommendation models
  • 32. Streamlining NLP model creation and inference
  • 33. Real-time, accuracy and lineage-aware featurization
  • 34. Making model cards
Question? Give us feedback → (opens in a new tab)Edit this page
Apply Conf 2022
11. How to draw an owl and build effective ML stack

11. How to draw an owl and build effective ML stacks, Sarah Catanzaro, Amplify

https://www.youtube.com/watch?v=wbExGrRBDvI&ab_channel=Tecton (opens in a new tab)

👉 Slides (opens in a new tab)

  • We need a common language to describe the key elements of the ML stacks and workflows

  • 3 layers

    Screen Shot 2022-05-24 at 11.36.47.png

    • Data management

      Screen Shot 2022-05-24 at 11.38.15.png

    • Model and deployment

      Screen Shot 2022-05-24 at 11.40.36.png

    • Model operations

      Screen Shot 2022-05-24 at 11.43.16.png

10. Effective system ML development12. Panel What engineers should know when building