MARS London

MARS (Mentorship for Alignment Research Students) is a research programme that pairs students with experienced mentors to work on an AI safety (technical or policy) research project for 2 to 3 months.

The programme will start with an in-person kick-off in London, although we are also open to remote applications, at the start of January and will continue until March, the projects will have a flexible start and end date depending on the project.

Applications are now closed, if you would like to express interest in our future programmes, then you can do so here.

Program Details

Our previous projects included AI Governance, Mechanistic Interpretability, Automated Alignment Research, and Dangerous Capabilities Evals.

Applications will close on 13th December at midnight. Depending on the projects you apply to, you may receive a coding test or be invited to an interview with the project mentors. We will aim to inform all applicants of the decision by Monday 18th December.

Timeline

  • 13th December - Applications close, late applications accepted until 15th.

  • 18th December - Decisions sent to attendees (some decisions may be slightly later depending on mentors).

  • 8th - 11th January - In-person kick-off in London.

  • 12th January - Project work with frequent collaboration with mentors and teammates.

  • ~22nd March - Estimated programme end date and research Symposium.

The end date will depend on the projects and the size of team, progress, and other requirements from mentors and participants.

This programme is being run in collaboration with Cambridge AI Safety Hub.

FAQs

Contact info@safeailondon.org if you have any questions