🎃 It's Hacktoberfest time! Join the virtual event on Oct 3 to learn how to contribute to open source. Register now

ML and LLM system design: 450 case studies to learn from

A database of 450 case studies from 100+ companies.

How do companies like Netflix, Airbnb, and Doordash apply AI to improve their products and processes? We put together a database of 450 case studies from 100+ companies that share practical ML use cases, including applications built with LLMs and Generative AI, and learnings from designing ML and LLM systems.

Navigation tips. You can play around with the database by filtering case studies by industry or ML use case. We added tags based on recurring themes. This is not a perfect or mutually exclusive division, but you can use the tags to quickly find:

Generative AI use cases. Look for tags “generative AI” and “LLM” to find examples of real-world LLM applications.
ML systems with different data types: computer vision (CV) or natural language processing (NLP).
ML systems for specific use cases. The most popular are recommender systems, search and ranking, and fraud detection.
We also labeled use cases where ML powers a specific user-facing "product feature": from grammatical error correction to generating outfit combinations.

Enjoy the reading! And if you find the database helpful, spread the word.

Maintained by the team behind Evidently, an open-source tool for ML and LLM evaluation and observability.

Real-world ML and LLM systems

All the content belongs to respective parties. We simply put the links together

Selection criteria. Here is how we selected the case studies to include:

It is a blog, paper, or article about a machine learning system or LLM application created in-house (not by a vendor that sells or implements ML solutions for others).
It has sufficient detail on the use case and implementation: who the application is for, the AI product design, evaluation criteria, deployment architecture, etc. The more, the better.
It covers a real-world ML or LLM system that is used in production.
It describes one particular ML or LLM use case. There are only a few exceptions that detail multiple projects at once — about causal ML or generative AI, for example.

Did we miss some great ML and LLM case studies? Let us know! Our Discord Community with 2500+ ML practitioners and AI Engineers is the best place to share feedback.

Get Started with AI Observability

Book a personalized 1:1 demo with our team or start a free 30-day trial.

Get demo

No credit card required