Machine Alignment Transparency & Security (MATS) Residency Program 2027

Applications are now open for the Machine Alignment Transparency & Security (MATS) Residency Program 2027. The programme is designed for experienced researchers who want to pursue independent, high-impact research focused on reducing catastrophic risks from advanced artificial intelligence.

The residency provides participants with the time, funding, mentorship, research infrastructure, and community needed to tackle long-term AI safety challenges. Residents work alongside leading AI safety researchers and advisors from frontier AI companies, AI safety institutes, and research organizations while pursuing their own research agenda.

About the Residency

The MATS Residency supports researchers who have already demonstrated strong research ability and are ready to lead independent projects rather than receive introductory research training.

Residents develop and execute original AI safety research with substantial autonomy while receiving structured milestone reviews and expert feedback. The programme focuses on foundational challenges such as AI alignment, oversight, transparency, evaluation, and the safety of increasingly capable AI systems.

The residency accepts up to 30 participants each year through multiple intake windows, with the Winter 2027 cohort beginning in early 2027.

Residency Benefits

Selected residents will receive:

• Annual compensation ranging from USD 155,000 to USD 285,000.
• Research compute with no preset cap, based on project needs.
• Visa sponsorship for international participants.
• Relocation assistance, where applicable.
• Office space in London, Berkeley, or Washington, DC.
• Flexible hybrid working arrangements.
• Medical, dental, vision, and life insurance (where applicable).
• Pension or retirement benefits depending on location.
• Dedicated funding to attend conferences, workshops, and research events.
• Access to experienced advisors from leading AI organizations.
• Membership in the global MATS research community.

Research Areas

Applicants may propose research related to:

• AI alignment.
• AI transparency.
• AI security.
• AI control.
• AI evaluations.
• AI governance and safety.
• Agentic AI systems.
• Multi-agent systems.
• Mechanistic interpretability.
• Other foundational AI safety topics that reduce catastrophic AI risks.

Residency Tracks

Applicants can apply under one of several research tracks, including:

Bridge Track

Designed for experienced professionals transitioning into AI safety research while building a strong AI safety research portfolio.

Incubation Track

Supports researchers interested in developing new research agendas, organizations, or innovative approaches to AI safety.

Anchor Track

A longer-term research pathway intended for experienced researchers pursuing ambitious AI safety projects over an extended period.

Eligibility Criteria

Applicants should:

• Demonstrate the ability to conduct independent, high-quality research.
• Have completed at least one substantial research-caliber project.
• Be capable of solving open-ended research problems using rigorous methods.
• Show strong motivation to contribute to AI alignment, transparency, or security.

Applicants do not need:

• A PhD.
• A formal research job title.
• A long publication record.
• Previous AI safety experience.

Research completed in academia, industry, government, policy, or other settings—including unpublished or confidential work—may be considered, provided applicants can demonstrate its quality and relevance.

Ideal Applicants

The programme is suitable for:

• AI researchers.
• Machine learning engineers.
• Software engineers.
• Research scientists.
• Technical policy researchers.
• Experienced professionals moving into AI safety.
• Independent researchers with demonstrated research capability.

Required Application Materials

Applicants should prepare:

• Online application form.
• Curriculum Vitae (maximum two pages).
• Contact information and current affiliation.
• Office location preference.
• Work authorization information.
• Highest academic qualification.
• One representative work sample.
• Track selection with a short explanation.
• Two written responses.
• A brief counterfactual response describing alternative plans if not selected.

Selection Process

Applications will be evaluated based on:

• Quality of previous research.
• Evidence of independent thinking.
• Technical rigor.
• Relevance of research experience.
• Potential to contribute to AI safety.
• Overall fit with the selected research track.

Shortlisted applicants may be invited to an interview or complete a paid work test during the first week of December 2026.

How to Apply

To apply:

1. Visit the official MATS Residency application portal.
2. Complete your profile and eligibility information.
3. Select the residency track that best matches your research interests.
4. Upload your CV and strongest research work sample.
5. Complete the required written responses.
6. Submit your application before the deadline.

Applicants will be notified of the outcome of their application by the end of November 2026, with successful candidates beginning the residency in early 2027.