nirdiamant/agents-towards-production

Agents Towards Production is an open-source collection of 28 code-first tutorials that guide you from building a basic AI agent prototype all the way to deploying it in a real enterprise environment. An "AI agent" is a program that uses a large language model to reason, plan, and take actions, like searching the web, reading documents, or calling APIs, rather than just answering a single question. The tutorials are delivered as Jupyter Notebooks (interactive coding documents) and cover the full journey: giving agents persistent memory using vector databases, connecting them to real-time web search, wrapping them in FastAPI web endpoints, deploying them inside Docker containers, adding security guardrails to prevent misuse, scaling to GPUs, coordinating multiple agents working together, and monitoring their behavior in production. Topics include frameworks like LangGraph and LangChain. You would use this resource if you have experimented with AI agents in a demo but don't know how to make them reliable and scalable enough for real users. It bridges the gap between a working notebook proof-of-concept and a system you can actually put in front of customers. The tech stack centers on Python, Jupyter Notebook, LangGraph, Docker, and FastAPI.