Full Program Details
The people building today's most capable AI systems keep asking the same hard question: how do you make sure a powerful model actually does what we intend? That question is called AI alignment, and the field is short of researchers who have the mathematics to work on it. Iliad runs a month long course designed to turn strong maths, physics and theoretical computer science students into exactly those researchers, and it charges no tuition.
The next round, the Iliad Intensive for January 2027, runs from 4 to 29 January 2027 in London or the San Francisco Bay Area. Applications close on 16 November 2026.
There is no tuition. Participants get a $5,000 travel and housing allowance, office space, and lunch and dinner five days a week.
Details at a glance:
- Organiser: Iliad
- Format: full-time, in person, five days a week from 10am to 6pm
- Dates: 4 to 29 January 2027
- Location: London, UK or the San Francisco Bay Area, USA
- Cost: no tuition
- Support: $5,000 USD travel and housing allowance, plus lunch and dinner on weekdays
- Application deadline: 16 November 2026
What You Will Learn
The curriculum is built from one day modules grouped into five clusters, and the exact content changes between rounds:
- Alignment: the core problem and how it is tackled in practice, including reward learning theory
- Learning: why deep learning works, training dynamics, singular learning theory and data attribution
- Abstractions, representations and interpretability: ML engineering and mechanistic interpretability, which studies what is happening inside a neural network
- Agency: decision theory, reinforcement learning, agent foundations and world models
- Safety guarantees and their limits: debate, backdoors and worst-case interpretability
The course materials, written by around 20 contributors, are lecture notes and worksheets, with pointers for further study. Coding sessions use Python, so knowing it helps, though Iliad says it is not strictly required.
Who Can Apply
- You have a strong mathematical background. Iliad says applicants typically are studying for, or have completed, a graduate degree in mathematics, physics or theoretical computer science.
- You do not need a PhD. Iliad looks for mathematical maturity, plus research experience, general ability and motivation.
- You will be 18 or older by the start date.
- You can commit to four weeks of full-time, in-person study in London or the Bay Area.
The application asks about your US and UK immigration status. If you are applying from Nigeria, Kenya, South Africa or another country that needs a visa for the UK or US, factor visa processing time into your plans and ask Iliad about your situation before you commit.
How to Apply
- Open the Iliad application form and select the January 2027 Intensive and your preferred location, London or the Bay Area.
- Upload your CV as a PDF.
- Answer the written questions: why you are applying, your career goals, the two hardest mathematical concepts you have mastered and why they were hard, the most useful alignment project you could work on, and which research area interests you most.
- State your US and UK immigration status and how you heard about Iliad.
- Submit before 16 November 2026 and look out for a confirmation email from admissions@iliad.ac.
GizPulse Verdict
Who this is for: Master's and PhD students or graduates in maths, physics or theoretical computer science, including strong mathematicians from Africa, who want a serious, tuition-free route into AI safety research.
Who should pause: If you are a beginner coder looking for a general AI or software bootcamp, this is not it. The course assumes advanced mathematics from day one.
One tip to stand out: The question about the two hardest concepts you have mastered is your proof of depth. Pick genuinely difficult ideas, explain precisely where you got stuck and how you got past it, and avoid textbook definitions.
See the dates, location options and application link on the official Iliad website.
