Articles

• How I Cracked the Meta Machine Learning Engineering Interview
  • Coding: 3 rounds = 6 questions. Medium Leetcode. Check out Leetcode top-10 Facebook-tagged questions.
    • Communicate clearly: (1) understanding (2) describe algo and complexity at a high level (3) implement (4) test
    • Bias for speed: During (2), bias for a sub-optimal solution that you can write quickly.
  • Design interviews: 2 rounds: system design, ML system design.
    • System design: solve a problem with a distributed system.
    • ML system design: solve a problem with ML, e.g. build a classification system or a rec system.

System Design

Tips

• Be proactive: you have to be the one to drive the conversation.
• Outline: (1) clarify the problem (2) propose a high-level design (3) draw out the design (4) zoom into certain components
• Practice drawing: Excalidraw.

Resources

• Educative’s Grokking the system design interview.
  • Lessons can be good/bad. Focus: REST API, load balancing, communication models.
• Educative’s Grokking the Machine Learning Interview.
• Alex Xu’s system design on YouTube channel ByteByteGo.
• Chip Huyen’s Designing Machine Learning Systems book. Website.
• Samuel Flender’s ebook Machine Learning on the Ground: Design and Operations of Real-World ML Applications. 50 pages.
• Research papers: Eugene Yan’s list of curated papers. Selected papers:
  • Amazon Search: The Joy of Ranking Products.
  • Deep Natural Language Processing for LinkedIn Search Systems.
  • Embedding-based Retrieval in Facebook Search.
  • Deep learning for recommender systems: A Netflix case study.
  • Wide & Deep Learning for Recommender Systems.
  • Deep Neural Networks for YouTube Recommendations.
  • Neural Collaborative Filtering.
  • 150 Successful Machine Learning Models: 6 Lessons Learned at Booking.com.