ARBOx
Alignment Research Bootcamp Oxford
A 2-week intensive designed to rapidly build skills in AI safety.
ARBOx5: 4–15 January 2027
Applications for ARBOx5 have now closed. To hear when the next round opens, register your interest.
- Length
- 2 weeks, full-time
- Where
- Trajan House, Oxford
- Cost
- Free for admitted participants
- Who
- Open to all, not just Oxford students
More info
ARBOx alumni have gone on to become MATS scholars, LASR participants, AI safety researchers and engineers at organisations like Anthropic, EleutherAI and Redwood Research, and to pursue PhDs at world-leading universities.
ARBOx5 is an in-person, full-time programme running from 4 to 15 January 2027 at Trajan House in Oxford. Last time, ARBOx4 brought together 29 participants from 29 June to 10 July 2026, ending with three-day capstone projects in mech interp, evals, RL and alignment science, alongside board games and movie nights, participant lightning talks and a trip to LISA in London.
Programme Details
We provide Central Oxford accommodation for non-Oxford residents and support with travel expenses, as well as lunch and dinner on weekdays. The programme is free for all admitted participants.
You do not need to be an Oxford student to participate (though we’d love to see a large number of applications from Oxford students!).
What you’ll do
During the programme, you’ll follow a curriculum specially made for ARBOx, adapted from the ARENA syllabus. You’ll cover the fundamentals of ML, build gpt-2-small from scratch, learn about mech interp, RL and evals, and complete a short capstone project by the end of the programme. There will be lectures in the morning covering aspects of the syllabus, and afternoon pair-programming with TA support. During lunch break, there will be short talks from experts in the field, and we will host socials in the evenings. By the end, you’ll have hands-on experience with ML and safety techniques, a concrete mini project, and clear next steps for contributing to AI safety.
Syllabus
We’ll likely cover:
- 01Learning the fundamentals of ML
- 02Building gpt-2-small from scratch
- 03An introduction to foundational mech-interp techniques, RL and evals
- 04Replicating some frontier alignment research (e.g. Emergent Misalignment)
- 05A short capstone research project
Who should apply?
We’re looking for applicants with basic familiarity with linear algebra, Python programming, and AI safety (e.g., having completed AI Safety Fundamentals or another fellowship). Application is competitive, though: many more applicants meet the technical requirements than we have places.
We’d encourage you to apply if you’re eager to work on technical AI safety and satisfy the technical prerequisites. The programme is designed to upskill participants in ML safety, targeting those who would benefit from this training, regardless of their background. If you have relevant graduate or professional experience, this is also a plus! It’s easy to underestimate your abilities and your potential to contribute to the field, and so we suggest you err on the side of applying.
Applications closed
Applications for ARBOx5 have closed. Register your interest to hear when the next round opens.
Get involved
If you’re new and interested in learning more about Oxford’s AI safety community, we’d love to have you sign up for our mailing list or join us at an upcoming event.
If you have questions about AI safety or OAISI, you’re also welcome to reach out — we’re happy to meet for a coffee, go for a walk, or have a quick call.