The Iliad Intensive is now accepting applications for its fully funded, four-week in-person programme in technical AI alignment research. Designed for researchers with strong backgrounds in mathematics, theoretical physics, computer science, and related disciplines, the programme provides advanced training in the foundations of AI safety and alignment.
Participants attend the Intensive in either London or Berkeley (Bay Area) and receive a US$5,000 travel and housing allowance, free tuition, office space, and weekday meals throughout the programme.
Programme Overview
The Iliad Intensive is a rigorous, full-time course that introduces participants to the core research areas shaping modern AI alignment. Rather than focusing on applied machine learning alone, the programme emphasizes theoretical foundations, interpretability, agent theory, and safety guarantees.
Programme: Iliad Intensive
Institution: Iliad
Format: In-person, full-time
Duration: 4 weeks (with occasional 3-week editions)
Schedule: Monday–Friday, 10:00 AM–6:00 PM
Locations: London (LISA) and Berkeley, California
Tuition: Free
Participant Support: US$5,000 travel and housing allowance
Additional Benefits: Office space, lunch and dinner five days per week
The programme runs monthly, alternating between London and the Bay Area.
What the Programme Covers
The curriculum is organized into five thematic clusters, with each module taught over a single day. Since there is more material than fits into one month, individual programme iterations vary slightly.
AI Alignment
Participants explore the alignment problem and its major research questions, including reward learning, model alignment during pre-training and post-training, deployment challenges, and different forms of objective misspecification.
Deep Learning Theory
The programme examines the mathematical foundations of deep learning through topics such as singular learning theory, training dynamics, and data attribution, helping researchers better understand why modern neural networks work.
Interpretability and Representations
This cluster introduces mechanistic interpretability, belief-state geometry, natural latent representations, computational mechanics, and practical machine learning engineering concepts.
Agency and Decision Theory
Researchers investigate reinforcement learning, idealized agency, AIXI, coherence, expected utility theory, world models, and theoretical approaches to modelling long-term goal-directed agents.
AI Safety Guarantees
The final cluster focuses on debate, complexity theory, backdoors, steganography, worst-case interpretability, and the limits of existing AI safety guarantees.
Who Should Apply?
The Iliad Intensive is intended for technically strong researchers who want to contribute to foundational AI alignment research.
Applicants are particularly encouraged if they:
- Are pursuing or have completed a graduate degree in mathematics, theoretical physics, computer science, or a closely related field.
- Possess a strong mathematical and analytical background.
- Demonstrate research experience or exceptional technical competence.
- Are motivated to pursue long-term research in AI alignment and AI safety.
A PhD is not required. The programme evaluates candidates based on research ability, technical potential, competence, and motivation rather than academic credentials alone.
Financial Support
The programme is fully funded, making it accessible to participants from around the world.
Successful applicants receive:
- No tuition fees
- US$5,000 travel and housing allowance
- Dedicated office workspace
- Lunch and dinner provided five days per week
- Access to Iliad’s research community and technical lectures
The allowance is intended to help participants cover travel and accommodation costs while attending the programme in London or Berkeley.
Why This Opportunity Is Valuable
AI alignment has emerged as one of the most important research fields in artificial intelligence, bringing together mathematics, theoretical computer science, machine learning, philosophy, and decision theory.
The Iliad Intensive offers participants a structured introduction to this interdisciplinary landscape while providing exposure to leading researchers and cutting-edge theoretical work. For graduate students, early-career researchers, and technically trained scientists interested in AI safety, it serves as a valuable pathway into advanced alignment research and future research fellowships.
How to Strengthen Your Application
A competitive application should emphasize technical depth rather than broad AI experience.
Applicants should highlight:
- Strong mathematical or theoretical coursework.
- Research experience in mathematics, physics, computer science, or machine learning.
- Evidence of independent problem-solving or research projects.
- Genuine motivation for pursuing foundational AI alignment research.
- Interest in theoretical rather than purely applied artificial intelligence.
Candidates should clearly explain why the programme aligns with their long-term research goals and demonstrate curiosity about topics such as interpretability, agent foundations, reinforcement learning, and AI safety.
How to Apply
Applications are submitted through Iliad’s official online application portal. The programme accepts participants for monthly in-person cohorts hosted alternately in London and Berkeley.
As the published programme information does not specify a fixed application deadline, interested candidates are encouraged to apply as early as possible for their preferred cohort.
Disclaimer
Global South Opportunities (GSO) is not the organisation offering this programme and does not participate in the admissions or selection process. Applicants should verify programme dates, eligibility requirements, and cohort availability through Iliad’s official website before applying.
