Agentic AI and LLM Engineering

The move from chat demos to working agents is where most teams stall. This programme takes you through how large language models actually behave, how to specify work for them, how to give them tools and grounded data, and how to evaluate, cost and guard them before they touch a customer. Every lesson is written in-house and assessed with practical work rather than recall.

10 weeks 60 guided hours Intermediate to Advanced

See membership

No account needed to buy. Pay with your card and we create your student account from the email you enter, with this programme ready inside it for twelve months.

Agentic AI and LLM Engineering course artwork

What you will achieve

Explain in plain terms what a language model can and cannot reliably do
Write prompts and schemas that behave like specifications, not wishes
Build an agent loop with tools, retrieval and memory
Evaluate agent output with a repeatable test set instead of vibes
Control cost, latency and failure modes before release

Entry guidance

— Comfortable writing basic Python or JavaScript
— Some experience calling an HTTP API

Original Mai Studies learning material. This is a professional completion programme, not a regulated qualification or university accreditation.

Programme syllabus

3 applied modules

Lessons: 6
Assessed tasks: 3
Practice questions: 9
Timed mock papers: 0

How language models actually behave

Build an accurate mental model of the technology before you build on top of it.

2 lessons · 3 practice questions · 1 assessed task

1. What the model is really doing36 min study · 3 learning objectives · practical activity
Tokens, context and probability, explained without hand-waving.
2. Prompting as specification21 min study · 3 learning objectives · practical activity
Move from chatty prompts to structured instructions with a schema and acceptance criteria.

Specify a model-backed task end to end — marked out of 100, pass at 80%

Building an agent that does real work

Tools, the agent loop, retrieval and memory, assembled into something useful.

2 lessons · 3 practice questions · 1 assessed task

1. Tools and the agent loop36 min study · 3 learning objectives · practical activity
How an agent decides to act, observes the result, and decides again.
2. Retrieval, grounding and memory21 min study · 3 learning objectives · practical activity
Give the agent real facts, and keep what matters between turns.

Design an agent for a real task — marked out of 100, pass at 80%

Evaluation, cost and shipping

Prove it works, know what it costs, and keep it safe in front of real users.

2 lessons · 3 practice questions · 1 assessed task

1. Evaluation you can defend21 min study · 3 learning objectives · practical activity
Build a test set and a scoring method instead of judging by impression.
2. Cost, latency and guardrails in production21 min study · 3 learning objectives · practical activity
What it costs per call, how slow it feels, and what stops it doing harm.

Production readiness review — marked out of 100, pass at 80%

Lesson pages, assessed tasks and mock papers open the moment you join this programme. Teaching material is studied online; the notes you write are yours to download.

What students say

No ratings yet.