By the end of day two, your developers have written their own agent.
A hands-on two-day workshop that takes developers from zero in Python to coding LLM-based agents. From language basics and core principles to building agents with local models, Claude and OpenAI, through the LangChain and LlamaIndex libraries and the Jupyter Notebook workflow — taught at the keyboard, by doing.
AI Agent Coding with Python — Hands-on Developer Workshop
The problem underneath.
There is a large gap between understanding agents conceptually and being able to build one. It is closed by writing code that fails, seeing why, and fixing it — which is exactly what a lecture cannot provide and what this workshop is entirely made of.
Two days, at the keyboard. Python fundamentals for those who need them, then straight into agent construction: tool use, chains, retrieval, and the practical differences between running against a local model, Claude and OpenAI. LangChain and LlamaIndex are taught as working libraries rather than surveyed, and the whole thing happens in Jupyter so the loop between trying something and seeing the result stays short.
What the program covers.
Introduction to Python & Core Principles
A from-scratch introduction to the language — data types, control flow and functions as a solid programming foundation.
LLM Agent Coding — Local · Claude · OpenAI
A hands-on walkthrough of end-to-end agent construction against local models, Claude and OpenAI.
LangChain & LlamaIndex
The two libraries used in practice for agent chains, tool use and data-driven RAG flows.
Development with Jupyter Notebook
Trial and error in an interactive notebook — cell-by-cell execution, rapid prototyping and sharing.
Who this is for — and who it isn't.
Most providers only answer the first half. The second half saves everybody a quarter.
A good fit if
- Your developers need to build agents and are learning by trial and error alone
- Your team is strong in another language but new to the Python AI ecosystem
- You want capability in-house rather than outsourced agent development
- Developers learn better at the keyboard than in a lecture
Probably not if
- Your team already ships production LLM agents
- You need business or executive-level understanding rather than coding ability
- You want the agents built for you — that is a consultancy engagement
Questions we get asked.
Do participants need Python experience?
No. The workshop starts from zero in Python. Developers experienced in other languages move through the first section quickly; the agent construction work is where everyone spends the time.
Which model providers are covered?
Local models, Claude and OpenAI — deliberately all three, because the practical differences between them (cost, latency, tool-use behaviour, data residency) matter more than any single API.
Is this the same as the Claude Code program?
No. That program changes how an existing engineering team works. This one teaches developers to build LLM agents from scratch in Python. Different audience, different outcome.
How many participants?
Small enough that everyone gets attention at the keyboard — typically up to twelve. Beyond that the hands-on quality drops sharply.
What do participants leave with?
Working agent code they wrote themselves, the notebooks, and a setup they can keep developing in.
Often combined with.
C-Level AI Consultancy
An end-to-end six-month consulting and training program designed for executives and decision-makers.
AI Model Watch and Stack Intelligence
A six-month analysis program that reviews and compares every cloud and local AI model in the world against your organization's real needs.
AI-Augmented Coding Consultancy
A six-day technical program that moves software teams to AI-assisted development.
Talk through AI Agent Coding with Python with us.
Start with a free 90-minute assessment. We map where this fits, what it would change, and what it would return — then you decide.