krishnakalyan3 / learning-ray

learning-ray

In this interview, you’ll talk with one of our Member of Technical Staff, Field Eng and undergo a live coding exercise. You’ll be asked to write a single-threaded algorithm and discuss how to scale it (with Ray). You should come prepared knowing Ray Core fundamentals and be familiar with Ray’s Python interface. We don’t expect you to have the APIs memorized–our engineer will help you along the way. However you should understand the fundamentals and some Ray best practices.

Introduction to Ray and Basic Concepts

TODO: Read Chapter 1-3 https://docs.ray.io/en/latest/ray-core/walkthrough.html

conda activate ray

Scaling Single-Threaded Algorithms

TODO: Read Chapter 6-7

Best Practices and Advanced Concepts

TODO: Read Chapter 11 https://docs.ray.io/en/latest/ray-core/tips-for-first-time.html

Machine Learning with Ray

Summary

• Install Ray and run basic examples.
• Understand Ray’s core concepts (Tasks and Actors).
• Write and explain a single-threaded algorithm.
• Parallelize the algorithm using Ray.
• Study Ray best practices and advanced features.
• Machine Learning with Ray

Links:

• 2023 Ray Summit
• Lessons From Fine-Tuning Llama-2
• Learning Ray
• Documentation