Agentic AI bootcamp
A practical Python programme for developers and technical teams learning to build AI agents. Guided exercises cover tool use, retrieval, evaluation and oversight, followed by a capstone project.
How we can help
Building an AI agent involves decisions about data, tools, control flow and what should happen when a step fails. This bootcamp takes participants through those decisions in Python. We introduce the underlying language model concepts, then work through exercises using frameworks such as LangChain and LangGraph. Each exercise adds a capability that participants can inspect and test.
The programme is suitable for developers, analysts and technical teams with working Python knowledge. Participants build towards a capstone based on an agreed use case, such as finding information in documents or assisting with a defined internal task. Alongside implementation, the course covers evaluation, limits on tool access and points where a person should review an action. The capstone provides a basis for further development; production requirements are reviewed separately.
When this may be useful
- Your team understands basic prompting and wants to learn how to build a multi-step application around a language model.
- You have an internal workflow suitable for a training project and want developers to explore its technical requirements together.
- Existing experiments need clearer evaluation, error handling and human review before you consider wider use.
What the work involves
Foundations and development setup
Work through language model inputs, structured instructions, tools, memory and control flow. Set up the Python environment and examine a small working example before extending it in the labs.
Retrieval and tool use
Build exercises that retrieve relevant information and call defined tools. Discuss how to choose inputs, inspect returned data and distinguish useful evidence from an incomplete or misleading response.
Evaluation and oversight
Create example tasks and expected behaviours, examine failures and add review points. Practise handling missing information, restricting tool actions and recording enough detail to understand what happened.
Capstone development
Apply the course material to an agreed task. Participants explain the design, demonstrate its behaviour and document the assumptions, limitations and further work needed for their capstone agent.
How the assignment runs
Confirm the starting point
Review Python experience, available tools and learning objectives. Select a capstone that fits the participants' level and the training scope.
Build through guided labs
Introduce concepts through short explanations and coding exercises, with time to inspect behaviour, troubleshoot errors and discuss design choices.
Test and present the capstone
Evaluate the completed workflow against example tasks. Review the implementation and agree the next learning or development steps.
What to prepare
These details will help us understand the starting point and agree a useful scope:
- An outline of participants' Python experience and development tools
- A proposed use case and sample data cleared for training use
- Details of available devices, connectivity and approved model or tool access
If some information is still being developed, we can discuss what is available in the first conversation.
Questions about this service.
For anything specific to your organisation, get in touch. We can discuss the requirements before you decide on an engagement.
How much Python do participants need?
Participants should be comfortable reading and modifying Python code and running scripts. Prior machine learning experience is not required. We review the group's starting level before agreeing the exercises.
Can the capstone use our organisation's documents?
Yes, where the documents are suitable and you have approved their use with the training tools. Anonymised or representative samples can be used when the original material is confidential.
Does completing the bootcamp make the agent ready for production?
The capstone demonstrates the skills covered in training. Production use may require further security, integration, performance and operational work, depending on the use case and your organisation's requirements.
Let’s talk about what you need.
Share a little about your project, the challenge you are facing and where you would like some help.