AI & Data Engineering

This is the technical track, built around the four things consultancies now hire for under "AI & Data": data engineering, machine learning, language models, and visualisation. Finance is moving off ad-hoc spreadsheets and onto real infrastructure — warehouses, version-controlled transformations, models that predict rather than just report. You work the way a modern data team works, and everything you build has to hold up when someone asks where the number came from.

Who it's for

Suits you if you would rather build the machine than run it by hand. It is the most technical track we run, and the closest to what AI and data roles actually advertise for. No programming background assumed — we start at the first SELECT.

What you'll cover

SQL and data engineering

Real queries against real tables, then the pipelines that keep them refreshed instead of rebuilt by hand each cycle.

Machine learning

Prediction and classification on problems that actually matter — churn, risk, forecasting — and knowing when a model is not the answer.

Language models

Putting LLMs to work on your own data: extracting structure from documents, and building tools people can ask questions of.

Power BI and visualisation

Past drag-and-drop into measures, relationships, and dashboards that survive contact with actual users.

What you're working toward

Take a raw source, model it, and ship something that runs on its own — a dashboard, a prediction, or an LLM tool — and be able to defend every number in it.

What you'll be building

Collect

Pull the numbers from wherever they actually live.

SQL · APIs

Clean

Shape and check them until they can be trusted.

Power Query · Python

Model

Structure the data so it can answer a question.

dbt · SQL

Predict

Learn from what happened to say what happens next.

scikit-learn · LLMs

Decide

Reporting someone can act on without asking you.

Power BI · Tableau

Pick a level to see what you actually build

Source connectionsIllustrative

Ingest

Market data APIConnected1.24M
ERP exportConnected340k
CSV dropSyncing12k
Manual uploadIdle—

Raw payload

01{ "date": "2024-11-04", "ticker": "EQNR", "close": 302.40 }
02{ "date": "2024-11-04", "ticker": "DNB",  "close": 214.80 }
03{ "date": "2024-11-05", "ticker": "EQNR", "close": 305.10 }
04{ "date": "2024-11-05", "ticker": "TEL",  "close": 138.25 }
05{ "date": "2024-11-06", "ticker": "DNB",  "close": 216.05 }

Sound like your track?

You pick a path once you're in — nothing here is binding, and plenty of members end up touching both.

Apply to join See the Excel & Analysis path