Tool
Pydantic
8 cases from 6 companies use Pydantic in their stack. 3 report results in production. Tasks: Evaluate models, Build an assistant over internal docs, Recommend to new users and new items, Learn user and item embeddings and 4 more. Methods: Structured output, Golden set / benchmark, LLM observability, Human in the loop, LLM-as-a-judge and 30 more. Companies: Raiffeisenbank, DoorDash, Etsy and 3 more.
The 52x Multiplier: How Zepto Mastered AI Agent Evaluation at Scale![]() | ![]() | e-commerce | 2026-06 |
| Building a unified consumer memory for personalization at scale | food-delivery | 2026-06 | |
From Chaos to Order: Extracting Structure from Unstructured Text with LLMs![]() | ![]() | marketplace | 2026-05 |
A Corporate AI Agent at Nornickel![]() | ![]() | general | 2026-05 |
| Understanding Etsy's Vast Inventory with LLMs | marketplace | 2025-10 | |
| Function Calling 101: One Agent Instead of a Zoo of Bots | — | general | 2025-05 |
| From Chaos to Control: Building a Streamlit Tracer for Debugging RAG Systems | general | 2025-05 | |
| Improving a Modular Chatbot with LLMs: Practical Experience | banking | 2025-05 |


