Autoscaling machine learning APIs in Python with Ray

If you have used web frameworks like Flask or Django to wrap a scikit-learn model when writing your own REST API from scratch, see how to create a much more performant application with fewer lines of code using Ray serve. https://github.com/jonathandinu/scaling-data-science

Additionally, this video covers the general process of deploying a machine learning model for production and best practices for building a scalable architecture.

⏭ next video: https://youtu.be/yLKHHiT2nWw
⏮ previous video: https://youtu.be/a051mbC9zqw
🎞 playlist: https://youtube.com/playlist?list=PLmetp36hFxeyc9qO_5tPNMW-YD3tZfCFN

🔗 Ray serve: https://docs.ray.io/en/latest/serve/index.html
🔗 Ray dashboard: https://docs.ray.io/en/master/ray-dashboard.html
🔗 Ray cluster: https://docs.ray.io/en/master/cluster/index.html
🔗 Locust: https://locust.io
🔗 scikit-learn: https://sklearn.org
🔗 Stack Exchange data explorer: https://data.stackexchange.com/stats/queries

Stay in the loop! 👇

https://discord.gg/nbyZ6EpUum
https://twitter.com/jonathandinu
https://jonathandinu.com
https://jonathans.courses

#python #machinelearning #datascience

✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎⚡︎✌︎

0:00 - Introduction
2:22 - Deploying a model
4:23 - Scalable architectures for serving models
9:07 - Building an API with Ray serve
12:00 - Machine learning in production with Ray Serve
16:17 - Load (or stress) testing our API with Locust
19:16 - Outro