The new way to learn AI.

Make the most of AI in the classroom while keeping learning at the centre: tutors that guide with hints instead of giving the answer, adaptive platforms to spot early what each student struggles with, materials and activities prepared in less time, assessment that stays reliable with AI around, and student data protected.

Practical project · Final project: a school dossier with a one-page policy, a pilot unit, a priced risk register, stop criteria and a 10-minute defence.

The model proposes. The chartered engineer signs. AI in BIM, design, site and inspection without delegating the calculation, the measurement or the report.

Practical project · Final project: an eight-sheet service dossier with zones and signatures, versioned data, channels with a human review tray, stop criteria and what gets counted on day 90, defended before project management.

The model flags, summarises and proposes. It does not block the account, close the incident or apply the patch.

Practical project · A service dossier: eight slides, ten minutes and what gets counted on day 90

A large ground-mounted solar plant, explained to someone who has never set foot in one. The model proposes. The person energises.

Practical project · A one-page note: day data, the 6 % announcement, what was not checked, preparation signed, energisation left blank

A fermented batch, told to someone who has never seen a fermenter. The model proposes. The learner neither doses nor releases.

Practical project · Final project: the batch sheet for lot L-0417, with the data, the 8 % feed proposal, what was not checked and an empty dosing signature, plus the next-day sheet that reads without the chat and is defended in five minutes.

A packing line, explained to someone who has never set foot in a plant. The model flags. The lab releases.

Practical project · Final project: the shift sheet for the case lot, with the data, the thickener proposal as a draft alert, what was not checked and an empty release signature, fileable without the chat and defended before quality, R&D and audit.

A small river, explained to someone who has never read a gauge. The model flags. A protocol role warns. You prepare the note.

Practical project · An afternoon note: the data point, the proposal, what was not checked, the warning signature left blank

A river with three possible sources, explained to someone who has never processed a discharge permit. The model apportions. A role signs. You prepare the note.

Practical project · Final project: Tuesday's note, with the data, the proposal, what was not checked (with the model's apportionment as a hypothesis) and the report's signature left blank, defended against six hard questions.

A wind farm, explained to someone who has never climbed a turbine. The model flags. A procedure role isolates. You prepare the note.

Practical project · A morning note: the data point, the proposal, what was not checked, the isolation signature left blank

Move from improvised prompts to structured, verifiable instruction systems with measurable quality.

Practical project · Final project: a versioned prompt system with an evaluation set and scoring rubric that rebuilds an existing prompt, measures the improvement against its baseline, includes a regression report for a model change and is defended before a decision-maker.

Design intelligent workflows and agentic systems that run reliably inside real operations.

Practical project · Final project: a working automated flow with guardrails and human checkpoints, with its go-live file, time saved measured against the baseline and a live stop-switch test in the defence.

Treat language, image, video and audio generation as a production capability, not a novelty: models chosen on criteria, reproducible parameters, review before release and provenance on every file.

Practical project · Capstone: a multimodal generation pipeline with quality, provenance and labelling controls, applied to a real brief and documented so another person can reproduce it.

Turn research, positioning and content production into an AI-assisted marketing system: grounded sources, reusable brand assets, compliance controls inside the workflow and measurement that separates what was produced from what actually sold.

Practical project · Capstone: an AI-assisted campaign system with a measurement plan, applied to your own campaign or to Lunara Bikes, with a prompt and claims library, an approval workflow, compliance controls and a report against the previous baseline.

Use CRM evidence, predictive signals and AI assistants to make better sales decisions.

Practical project · Final project: a pipeline score validated against closed history and a forecast review with explicit assumptions, confidence range and measured error, built on a real CRM structure, with its limits, go-live plan and governance sheet, defended before a committee.

Understand the models, data and evaluation concepts behind modern Artificial Intelligence systems.

Practical project · Final project: a model capability and evaluation dossier for the four Albarén Seguros proposals.

Redesign knowledge work with reliable AI assistance, verification and measurable time savings.

Practical project · Final project: a verified, measured personal AI work system, with three documented workflows, their baseline, their verification protocol, their data limits and before-and-after measurement.

Design agentic systems that reason over goals, use tools, retain context and escalate safely.

Practical project · Agent pilot dossier: a level-1 or narrow level-2 refund agent, with a one-page sheet, allow-listed tools, logging, an owner and a stop.

Build trustworthy AI systems through evaluation, privacy, security, fairness and meaningful human oversight, and prove it with a case file someone else can review.

Practical project · Final project: an AI assurance case file for a real use case, with impact assessment, evaluation evidence, red-teaming report, human-oversight design, monitoring and incident plan, and a signed decision.