Machine Learning System Design

While preparing for job interviews I found some great resources on Machine Learning System designs from Facebook, Twitter, Google, Airbnb, Uber, Instagram, Netflix, AWS and Spotify.

I find this to be a fascinating topic because it’s something not often covered in online courses.

Twitter

• Using Deep Learning at Scale in Twitter’s Timelines
• Improving engagement on digital ads with delayed feedback
• Embeddings@Twitter

Instagram

• Lessons Learned at Instagram Stories and Feed Machine Learning
• Powered by AI: Instagram’s Explore recommender system

Facebook

• Deep Entity Classification: An abusive account detection framework
• New progress in using AI to detect harmful content

Uber Eats

• Food Discovery with Uber Eats: Building a Query Understanding Engine
• Food Discovery with Uber Eats: Recommending for the Marketplace
• Food Discovery with Uber Eats: Using Graph Learning to Power Recommendations

Uber

• Applying Customer Feedback: How NLP & Deep Learning Improve Uber’s Maps
• Forecasting at Uber: An Introduction

Airbnb

• Using Machine Learning to Predict Value of Homes On Airbnb
• Listing Embeddings in Search Ranking
• Learning Market Dynamics for Optimal Pricing
• Categorizing Listing Photos at Airbnb
• Applying Deep Learning To Airbnb Search
• Discovering and Classifying In-app Message Intent at Airbnb

Airbnb Experiences

• Machine Learning-Powered Search Ranking of Airbnb Experiences

Linkedin

• An Introduction to AI at LinkedIn
• Fairness, Privacy, and Transparency by Design in AI/ML Systems
• Communities AI: Building Communities Around Interests on LinkedIn
• Preventing abuse using unsupervised learning

Google

• Jeff Dean On Large-Scale Deep Learning At Google

Netflix

• A Multi-Armed Bandit Framework for Recommendations at Netflix

Spotify

• For Your Ears Only: Personalizing Spotify Home with Machine Learning
• How Does Spotify Know You So Well?

In addition, here are some resources on a more general process. Starting with the book Data Science for Business which explains the CRISP-DM (Cross Industry Standard Process for Data Mining).

The process involves six stages:

1. Business Understanding
2. Data Understanding
3. Data Preparation
4. Modelling
5. Evaluation
6. Deployment

Here is a more high-level breakdown on how to apply CRISP-DM on AWS.

Facebook also created a video series where they go into depth in how they structure Machine Learning Projects with the Facebook Field Guide to Machine Learning.

I hope you found this useful! For more content, follow me on Linkedin.