Applied AI2 min read
Evaluating LLM features before you ship them
Demos are easy; reliable AI features are not. How to build a small evaluation set, choose criteria, and decide whether a feature is ready.
Applied AI programme
Build, evaluate and ship reliable LLM features: prompting, RAG, agents and safety.
Who it's for
The Applied Generative AI Bootcamp is for developers, analysts and product professionals with basic Python who want to build real features on large language models. You will learn how LLMs behave, write and test prompts, build retrieval-augmented generation, use tools and agents, evaluate quality, manage safety and cost, and ship a working AI feature reviewed by practitioners.
Curriculum
9 modules, each ending in a piece of real work you keep for your portfolio.
Build an accurate mental model of what LLMs do, so you can predict their strengths and failure modes.
Call LLMs from Python reliably, handle errors and get outputs your code can depend on.
You'll produce: Command-line assistant that returns validated structured output
Write prompts as engineered artefacts: versioned, tested and improved against evidence.
Measure quality before you ship, so changes to prompts or models are decisions rather than guesses.
You'll produce: Evaluation harness for a prompt-based feature
Ground model answers in your own documents, with citations users can check.
You'll produce: RAG assistant over a document collection with source citations
Let models take actions through tools, while keeping their behaviour predictable and bounded.
Protect users and organisations from the specific risks that LLM features introduce.
Move from a notebook to a small service that others can use, monitor and improve.
You'll produce: Deployed AI application with logging and an evaluation report
Decide which problems are suited to generative AI and design experiences that earn user trust.
Schedule
The live-class schedule is published with each cohort. Your advisor will share it before you enrol.
Interactive classes with practitioners
Every module ends in a deliverable
Written reviews on your work
Critique and study with your cohort
Mentorship
Each cohort is paired with engineers and practitioners who build LLM applications, and who review your code, evaluation approach and design decisions.
Small-group build sessions where mentors debug pipelines and discuss design choices with you.
Scheduled one-to-one sessions to scope your capstone and plan how to apply AI in your current or target role.
Code reviews on your repositories with specific feedback on reliability, safety and structure.
Capstone reviews that test whether your evaluation results support the claims you make about your feature.
Career support
Preparation, not promises — we don’t guarantee jobs; hiring decisions belong to employers.
Present your repositories and demos so that reviewers can see what you built, how you tested it and its limits.
Describe your AI projects accurately and specifically, without overstating what they do.
Practise explaining LLM system design, RAG trade-offs and evaluation choices, with feedback.
For working professionals, prepare a proposal to introduce an AI feature or workflow in your current organisation.
Prepare a concise summary and demo links so you can respond quickly when a referral opportunity appears.
Follow AI and engineering roles on the ActuateMinds job board and plan applications around your project work.
Have more questions or need more information?
Request a call backPortfolio
Portfolio-ready work, reviewed by mentors.
Build a service that classifies incoming support tickets, extracts key fields into a structured format and drafts a first reply. Evaluate accuracy against a labelled set and report where it fails.
Create a RAG assistant that answers questions about a set of public policy or HR documents with citations. Measure retrieval quality and handle questions the documents cannot answer.
Build an agent that can query a sample database through a safe SQL tool and summarise the result. Restrict its permissions, log every action and test it against prompt injection attempts.
Identify a real problem in your work or community, build an AI feature to address it and deploy it with logging and evaluation. Present the product case, quality results and known limits to a mentor panel.
Toolkit
Certificate of completion
Your Name
has completed the required projects of the Applied Generative AI Bootcamp.
Certificate
Awarded once you complete the programme's required projects. The reviewed projects themselves are what you'll show employers.
Timeline
Module order may be refined between cohorts. Week-by-week dates are shared with each cohort.
Kick-off
Dates are confirmed per cohort.
Module 1: How large language models work
Build an accurate mental model of what LLMs do, so you can predict their strengths and failure modes.
Module 2: Working with model APIs
Produce: Command-line assistant that returns validated structured output
Module 3: Prompt engineering
Write prompts as engineered artefacts: versioned, tested and improved against evidence.
Module 4: Evaluating LLM applications
Produce: Evaluation harness for a prompt-based feature
Module 5: Retrieval-augmented generation
Produce: RAG assistant over a document collection with source citations
Module 6: Tool use and agents
Let models take actions through tools, while keeping their behaviour predictable and bounded.
Module 7: Safety, privacy and responsible use
Protect users and organisations from the specific risks that LLM features introduce.
Module 8: Shipping an AI feature
Produce: Deployed AI application with logging and an evaluation report
Module 9: AI product thinking
Decide which problems are suited to generative AI and design experiences that earn user trust.
Wrap-up
Complete the required projects to earn your certificate.
Admissions
Apply online
Tell us about your technical background, current role and what you want to build with generative AI.
Advisor conversation
Discuss your goals, the schedule and fees with an admissions advisor. Fees are shared during this conversation.
Technical fit check
Complete a short Python exercise so we can confirm you are ready for the pace of the bootcamp.
Offer and onboarding
Accept your offer, complete enrolment and work through pre-work covering environment setup and API basics.
FAQs
Fees are shared during your advisor conversation, where we also explain what is included and any payment options available. We publish fees on this page once they are confirmed for the cohort. Separately, you may incur small usage charges from model providers when you run your projects, and your advisor will explain how to keep these low.
The bootcamp combines live sessions with hands-on building between them. Session timings and the time commitment you should plan for are confirmed per cohort. Your advisor will go through them with you. Because the work is project-based, expect to spend focused time coding and debugging outside live sessions, not only attending them.
Whether sessions are recorded, and how catch-up material is provided, is confirmed per cohort and explained in your advisor conversation. Live build sessions are where mentors help you debug and reason through design choices, which is hard to get from a recording. We recommend planning to attend live and treating any catch-up material as a backup.
You should be able to write and run basic Python scripts, including functions, lists and dictionaries, and install packages. Familiarity with APIs, JSON and Git is helpful. You do not need a machine learning background. If you are new to Python, build that foundation first; the technical fit check helps us confirm you are ready.
It will help you build demonstrable skills and projects, but we do not guarantee jobs or titles. Career support includes portfolio and resume reviews, mock technical interviews, referral readiness and job board access. Many learners use the bootcamp to bring AI into their current role, and mentors can help you plan either path realistically.
Yes. A certificate of completion is issued when you complete the required projects, including the capstone. For AI roles, a working, well-evaluated project that you can explain in depth usually says more than a certificate, so the bootcamp is designed around building and documenting real applications you can share with employers or your team.
You will work with hosted model APIs such as those from OpenAI and Anthropic, and frameworks such as LangChain and LlamaIndex, alongside plain Python so you understand what the frameworks do underneath. The field changes quickly, so the curriculum focuses on principles like evaluation, retrieval and safety that remain useful as specific models and libraries change.
Refunds are covered by our published refund policy at /refund-policy, which sets out the conditions and timelines. Please read it before enrolling. If you have questions about how it applies to you, your admissions advisor can explain the details for your cohort so you can make an informed decision before you pay.
Request a call back for personalised guidance on fit, schedule and fee.
Request a call backTalk to an advisor
Leave your details and an advisor will walk you through the schedule, fee and fit. No obligation.
Applied AI2 min read
Demos are easy; reliable AI features are not. How to build a small evaluation set, choose criteria, and decide whether a feature is ready.
Applied AI2 min read
Better prompts help, but reliable AI features come from product design: narrowing the task, grounding the model, and planning for failure.