Course · Starts February 22, 2026
Context Engineering for Football
A unique chance to get ahead of the game — learn how to build AI analysts for football using the methods we have developed at Twelve. Invite-only, for highly talented individuals willing to dedicate 10 hours a week.
Overview
During the last two years, Twelve Football have developed a unique approach to using LLMs and building AI agents. When we started working on our approach it didn't have a name; more recently it has come to be referred to as context engineering. The central technical idea is to build a data context for any question a user poses and then transform the data into insight.
From a footballing perspective this allows us to build AI analysts that give insight into scouting, match analysis, set pieces and more. This is the basis of our Earpiece product, which has been rapidly adopted by clubs and national teams.
Two types of people are suited to this course: those talented in understanding and writing about the game with some knowledge of Python, and those who already work in data science or engineering and have some knowledge of football. You'll get full access to a sandbox version of our tooling.
Due to very high demand we have closed applications for this cohort. Participation is now invite only. From Autumn 2026 this course will be part of our paid course package.
What you'll leave with
- Understand how LLMs work, how they're trained, prompted and turned into agents
- Wordalise football data — combining statistical and normative models with visualisation
- Combine human and machine understanding in analyst-style outputs
- Apply machine learning models and automated data analysis to football
- Use function calling, embedding search and information organisation to give models the right context
- Plan an analyst, build storyboards and improve interaction
- Ship two projects: wordalising football data and building an AI analyst
Curriculum
Module 1
Week 1 — Intro
How LLMs work: training, prompting, agents and the building blocks of context engineering.
Module 2
Week 2 — Wordalisation 1
Statistical and normative models, visualisation and wordalisation, and model cards.
Module 3
Week 3 — Wordalisation 2
Combining human and machine understanding in analyst-grade outputs.
Module 4
Week 4 — Wordalisation 3
Machine learning models and automated data analysis applied to football.
Module 5
Week 5 — Context 1
Function calling, embedding search and organising information for retrieval.
Module 6
Week 6 — Context 2
Planning an analyst, creating storyboards and improving interaction.
Module 7
Week 7 — Context 3
Putting it all together into a working AI analyst.
Module 8
Week 8 — Future
Reasoning, and what LLMs can and can't do today.
Course leaders
Professor David Sumpter
Co-founder, Twelve Football
Professor David Sumpter is a professor of applied mathematics at Uppsala University and the author of Soccermatics, a book that explores the application of mathematical modelling to football. He co-founded Twelve Football and has worked with major clubs and national federations worldwide.
Ágúst Pálmason Morthens
Lead Data Scientist, Twelve Football
Ágúst Pálmason Morthens is Lead Data Scientist at Twelve Football. He works on a daily basis with clubs to help them become more data-driven, translating football data into practical insights for decision-making.
Who it's for
- People working in football with some Python skills (Streamlit, GitHub)
- Scouts or match analysts producing written reports
- Football data scientists, or Soccermatics graduates
- Engineers doing technical work with LLMs who want to apply them to sport
Testimonials
"The course teacher dedicated time to sit with me individually, start from scratch, and write a customised step-by-step plan. This was going above and beyond and really assisted me as a relative beginner in Python."
Liam Cody
Pro Solutions Team Lead, StatsPerform
"Thank you for all this course does for the community. I'll be joining a Brazilian club as a data scientist in the upcoming weeks, and it is crystal clear that it would not happen without your course."
Rodrigo Salvador
Red Bull Bragantino
"I appreciated the course very much. I recommend it for anyone interested in gaining additional knowledge in data science and analysis in their organization."
Martin Vogelbein
Scouting, Match Analysis — Deutscher Fußball-Bund
"Especially the insight into presenting the results via Python Streamlit was a big game changer for me. We are now establishing this framework at TSG Hoffenheim."
Fabian Rupp
Business Intelligence & Data Analytics, Hoffenheim
Get the latest on football data and AI
Get sharp insight on how the game is changing, practical ideas you can use in your work, and early access to new features and updates from Twelve.