Work with AI models
Send requests, understand responses, and use results in an application.
Instructor-Led Engineering Bootcamp
Agentic AI Engineering Bootcamp
Learn to build AI applications that do more than answer questions.
AI can answer a question. But what happens when an application needs to find information in company documents, check a live system, complete a task, or ask a person for approval?
That is what you will learn in this hands-on bootcamp. Starting with the fundamentals, you will see how AI models, knowledge, software tools, and workflows come together to create useful AI-powered applications.
One practical project grows throughout the course. By the end, you will have worked on an Enterprise Service Agent and learned how to explain its architecture, test its behaviour, and identify where human oversight is necessary.
- 3-week bootcamp
- 15 instructor-led sessions
- 90 minutes per session
- One evolving project
No previous AI experience or Python knowledge required. Basic programming knowledge in any language is recommended and expected.
Why this bootcamp?
Go beyond prompts. Understand how real AI applications work.
Getting a response from an AI model is only the beginning. A useful AI application must also know which information to trust, when to use a tool, how to handle mistakes, and when to stop or involve a person.
Connect AI to knowledge
Let an application find relevant information in documents instead of relying only on what a model remembers.
Use tools responsibly
Connect AI-driven workflows to services such as a ticketing system or user directory.
Build controlled agents
Create workflows that can choose a next step, while your software enforces clear limits.
Check quality and safety
Test results, investigate failures, and decide which actions should require human approval.
The goal is not to memorize a framework. It is to understand the engineering decisions behind a working AI system.
Who Should Join?
Who is this bootcamp for?
This bootcamp is designed for people with basic programming knowledge who want a practical introduction to AI engineering.
Software developers
Learn how to add LLMs, retrieval, tools, and agent workflows to applications.
QA and test engineers
Learn to evaluate systems whose responses may change and to design meaningful failure tests.
IT, support, and operations engineers
Explore how familiar service workflows can be improved with knowledge retrieval and controlled automation.
Final-year students and early-career engineers
Build hands-on experience with the main building blocks of an AI application.
Technical professionals moving into AI
Understand the concepts, architecture, and coding patterns needed to begin AI engineering work.
What you need before joining
You should be comfortable with basic programming concepts such as variables, conditions, loops, and functions in any language.
You do not need prior experience with Python, machine learning, large language models, LangGraph, RAG, or MCP. We introduce the Python concepts used in the course before building the AI application.
This is a technical, code-based bootcamp, not a no-code course. You do not need advanced coding experience, but you should be willing to read and modify small pieces of code.
What You Will Learn
Six Practical Modules
Start with Python and the foundations of AI applications, then learn how to connect knowledge, tools and agents into a working system.
Module 00
Setup and Python Essentials
Get comfortable with the course environment and the Python features needed for AI applications. Learn to work with APIs, validate data, handle errors, manage configuration, and follow asynchronous code used in later examples.
What you will be able to do
You can run the prepared project, understand its structure, and make small changes confidently.
Your Project at This Stage
The Enterprise Service Agent is running, connected to an AI model, and ready to extend.
Module 01
Large Language Model (LLM) Foundations
Understand how applications communicate with AI models. Explore prompts, messages, tokens, context limits, structured outputs, and tool calling. Learn why models sometimes give incorrect or inconsistent answers, and why software must check important results.
What you will be able to do
You can use model responses inside an application rather than treating them as plain chat messages.
Your Project at This Stage
The agent can interpret a service request and return information in a structured format your code can check.
Module 02
Context Engineering and RAG
Learn how to help an AI application answer questions using your own documents. You will work with document chunks, embeddings, a searchable knowledge store, and retrieval-augmented generation (RAG). You will also explore citations, access permissions, and common retrieval failures.
What you will be able to do
You can build and troubleshoot an AI-powered document-search flow that uses relevant source material.
Your Project at This Stage
The agent can find relevant company guidance and support its answers with retrieved sources.
Module 03
AI Agents and LangGraph
Learn the difference between a predefined workflow and an agent that can select its next step. Explore common agent patterns, workflow state, routing, stopping conditions, and human checkpoints. Use LangGraph to connect the steps into a controlled flow.
What you will be able to do
You can explain and build a basic agent workflow that makes bounded decisions and knows when to stop.
Your Project at This Stage
The agent can decide whether to retrieve information, request a tool action, or hand control back to a person.
Module 04
Multi-Agent Systems and MCP
Understand when one agent is enough and when specialist agents may help. Learn how agents hand off work and how the Model Context Protocol (MCP) provides a common way to expose tools and other capabilities to AI applications. Explore the importance of safe access to enterprise systems.
What you will be able to do
You can design a simple specialist-agent workflow and understand how to connect tools through MCP.
Your Project at This Stage
Specialist capabilities can work together, and tools can connect the agent to enterprise services through controlled interfaces.
Module 05
From Manual Processes to Production-Minded AI
Start with a real business process and identify which steps need conventional code, an AI model, retrieved knowledge, a tool, an agent, or a person. Introduce evaluation, debugging, security safeguards, cost awareness, and human approvals. Bring these ideas together in the final project review.
What you will be able to do
You can assess whether an agent is appropriate, identify key risks, and explain how you would test the system before wider use.
Your Project at This Stage
You review the end-to-end system, demonstrate a targeted improvement, and test how it behaves when things go wrong.
How You Will Learn
Understand the system first. Then work with the code.
The bootcamp is instructor-led and architecture-first. Each session begins with a familiar problem, explains the underlying idea, and shows how that idea fits into the Enterprise Service Agent.
A typical 90-minute instructor-led class
5m
Concepts · 20 min20m
Architecture · 15 min15m
Live Coding + Walkthroughs · 40 min40m
Review · 10 min10m
Connect to the problem
Review the previous lesson and the problem to solve.
Understand the concept
Learn the idea in plain language before going into implementation.
Build the architecture on a whiteboard
See how the components fit together, one step at a time.
Live coding and code walkthroughs
Follow the instructor through working code, watch a focused change coded live, and practise making a similar change in the prepared project.
Run, inspect, and review
Execute the flow, examine results, and discuss what could go wrong.
During the code walkthroughs, you will understand how the pieces connect. During live coding, the instructor will demonstrate a focused implementation and explain the decisions along the way. You will also have guided opportunities to work with the code yourself.
You will not need to build an entire AI system from a blank editor in one class. We start with prepared code so the time is spent on understanding, practising, and learning from the results. Session code, architecture explanations, and exercises are provided through GitHub.
The Project You Will Build
Meet your Enterprise Service Agent
Imagine an employee says:
“I need VPN access.”
A basic chatbot might give general instructions. An Enterprise Service Agent can go further: understand the request, find the relevant company guidance, check information from connected systems, and help create a support ticket, with the right permissions and human approval when needed.
01Understand an employee request
02Find the relevant company guidance
03Check information from connected systems
04Help create a support ticket with appropriate controls
This is the one practical project you will work on throughout the bootcamp. As you learn new concepts, you will add capabilities to the same application instead of starting unrelated exercises.
You will start by running and understanding the application, gradually make it smarter and more useful, and finish by testing and explaining the decisions behind it.
The six modules you have just explored come together in this one practical application.
Your Final Project
Show what you can build and explain why it works
Your capstone builds on the Enterprise Service Agent developed throughout the bootcamp. You will make a targeted improvement and explain how it fits into the wider solution.
Capstone Project
Improve and explain your Enterprise Service Agent
Your capstone builds on the Enterprise Service Agent developed throughout the bootcamp. You will make a targeted improvement and explain how it fits into the wider solution.
The emphasis is on practical understanding. A successful demo matters, but so does knowing what could fail and how to detect it.
Your project review will cover
- A working change: Demonstrate the feature or improvement you made.
- A simple architecture explanation: Show how the main components work together.
- The right level of automation: Explain what should use code, AI, retrieval, tools, agents, or human review.
- Safety and permissions: Identify actions that need additional controls or approval.
- A testing plan: Show how you would check both normal and failure scenarios.
Tools and Technologies
Technologies You Will Work With
We introduce the tools through practical examples so that you understand what each one does and why it is useful.
Python
The programming language used for the application and its tools.
Pydantic
Helps check that data has the expected structure.
LLM APIs
Let applications send requests to AI models and receive responses.
Embeddings and vector search
Make it possible to find documents by meaning, not only exact keywords.
RAG
Adds relevant information from documents to an AI request.
LangGraph
Organizes the steps, decisions, and state of agent workflows.
MCP
A standard way to connect AI applications with external tools and capabilities.
Evaluation and tracing
Help you check quality and understand what happened during execution.
You will receive access to a model API for the guided course exercises, as described in the course plan.
Course Format
Bootcamp Format at a Glance
Instructor-led learning with guided practice and one evolving project.
| Course detail | What to expect |
|---|---|
| Duration | 3 weeks |
| Sessions | 15 instructor-led sessions |
| Session length | 90 minutes |
| Total learning time | 22.5 hours of live instruction |
| Learning style | Whiteboard architecture, live coding, code walkthroughs, guided exercises, and project reviews |
| Prerequisite | Basic programming knowledge in any language |
| Python / AI experience | Not required before joining |
| Project | One evolving Enterprise Service Agent |
| Materials | GitHub-based code, architecture notes, and exercises |
| Delivery | Cohort-based, instructor-led |
FAQ
Frequently Asked Questions
Answers to common questions about prerequisites, learning format and the practical project.
Do I need to know Python?
No. We introduce the Python features used in the bootcamp. However, you should already understand basic programming concepts in at least one language.
Do I need experience with AI or machine learning?
No. The LLM module introduces the concepts from the beginning and connects them to familiar software ideas.
Is this suitable for students or people early in their careers?
Yes, if you have basic programming experience. The lessons combine familiar examples, live coding, code walkthroughs, and guided practice. You do not need prior experience building AI systems.
Is the bootcamp useful for testers and operations engineers?
Yes. You will learn how to trace AI workflows, check outputs, identify failure scenarios, and understand where controlled automation can help operational processes. The coding prerequisite still applies.
Will I be writing code?
Yes. Yes. Sessions include instructor-led live coding, code walkthroughs, and guided practice with prepared projects. You will see the code being developed, understand how it works, and make focused changes yourself.
Do I need my own model API key or cloud account?
The course plan includes model API access for guided exercises. Separate cloud infrastructure is not part of the bootcamp requirements described here.
Is this a complete production deployment course?
No. This is a focused bootcamp that introduces the foundations of building and evaluating AI-powered applications. Large-scale deployment, cloud operations, model training, and research-level RAG techniques are outside its main scope.
Will I have a project to discuss or demonstrate?
Yes. You will work with the Enterprise Service Agent throughout the course and complete a capstone improvement supported by an architecture explanation and testing approach.
Can the bootcamp be delivered for a private company team?
Yes. The course plan supports private cohorts, and the reference case study can be adapted to a relevant internal business process.
Start Your AI Engineering Journey
Ready to start building with AI?
Learn the foundations, work on one practical project, and develop the engineering judgment needed to move beyond a basic chatbot demo.