smolagents

An in-depth exploration of the engineering principles learned from Hugging Face's Shippy project, focusing on code-first agents, smolagents, and robust LLM orchestration.

Discover how to move beyond fragile, single-pass Text-to-SQL pipelines by using Hugging Face smolagents. This guide explores the ReAct framework, CodeAgent implementations, and how to build self-correcting database agents using high-performance LLMs.

Discover how to build DeepMath, a high-performance math reasoning agent using Hugging Face's smolagents library and n1n.ai APIs. Learn why code-as-actions is the future of LLM reasoning.