The MATS Residency Winter 2027 is inviting talented individuals from around the world to apply for its upcoming intake, offering an opportunity to pursue ambitious research and safety-relevant projects focused on reducing catastrophic risks from advanced artificial intelligence.

The Residency is designed for people who can demonstrate strong research ability and the potential to produce meaningful work in AI safety. Importantly, applicants are not required to have a PhD, a formal “researcher” job title, an extensive publication record, or previous AI safety experience.

The Winter 2027 application window runs from 25 August to 31 October 2026, giving prospective applicants several weeks to prepare and submit their applications.

The MATS Residency is intended to support individuals working on important questions surrounding AI alignment, AI safety, AI governance, evaluations, assurance and advanced AI systems.

Rather than relying primarily on conventional academic credentials, the program evaluates the quality, seriousness and potential impact of an applicant’s work.

This makes the Residency particularly relevant to talented researchers, technical professionals, policy experts, entrepreneurs and individuals transitioning into AI safety who can demonstrate that they have the ability to tackle difficult, open-ended problems.

The program is also internationally accessible. Applicants can express a preference for one of three office locations — London, Berkeley or Washington, DC — and the Residency provides visa support for participants who need to relocate internationally.

Apply for the MATS Residency Winter 2027

Opportunity Category

Category: Fellowship / Research Residency

The MATS Residency is a structured research program that supports participants in developing AI safety research, safety-relevant projects or incubation initiatives.

MATS Residency Winter 2027 Deadline

The application window for the Winter 2027 intake is:

Opening date: 25 August 2026
Application deadline: 31 October 2026
Deadline timezone: Anywhere on Earth (AoE)

Applicants should avoid waiting until the final day because the application requires several components, including references, a work sample and written responses.

The next application window after the Winter 2027 intake is scheduled to open in March 2027.

When Will Applicants Hear Back?

Applications are reviewed by the MATS Residency team, with input from program advisors where appropriate.

The expected selection timeline is:

1. Applications close: 31 October 2026.
2. Application review: Applications are assessed by the Residency team.
3. Application outcome: Applicants are expected to hear about the outcome by the end of November 2026.
4. Interviews or paid work tests: Shortlisted applicants may be invited to an interview or paid work test during the first week of December 2026, depending on their track and research area.
5. Final selection: Selected Residents are expected to be informed by the end of 2026.
6. Residency start: Successful applicants are expected to begin in early 2027.

Information about interviews and panel members will be provided to shortlisted applicants beforehand.

Who Should Apply for the MATS Residency?

One of the most important features of the MATS Residency is its approach to evaluating applicants.

The program emphasizes what an applicant can do and the quality of their work, rather than simply looking at conventional credentials.

You do not necessarily need:

• A PhD.
• A master’s degree in AI safety.
• A job title containing the word “researcher.”
• A long list of academic publications.
• Previous professional experience specifically in AI safety.
• A conventional academic research career.

The program is interested in talented individuals who can demonstrate the ability to produce outstanding AI safety research or safety-relevant work.

This can include people whose strongest work is:

• Confidential.
• Unpublished.
• Conducted outside academia.
• Completed as part of an industry team.
• Based on technical or policy work.
• Demonstrated through projects rather than traditional publications.

If an applicant is uncertain about whether their background qualifies, MATS encourages them to apply or contact the Residency team before submitting an application.

What Research Does MATS Want to Support?

Applicants are expected to submit a brief plan describing what they would like to accomplish during the Residency.

Depending on the chosen track, this may take the form of:

• A research plan.
• A plan for safety-relevant work.
• A project plan.
• A plan to build or test an idea.
• A business or organizational plan for an incubation project.

The proposed work should explain its theory of change — in other words, how the proposed research or project could ultimately contribute to improving AI safety and reducing catastrophic risks associated with advanced AI.

1. Impact

MATS places significant emphasis on potentially high-impact work.

The Residency is particularly interested in ambitious, longer-term projects that may not fit neatly within conventional academic or industry structures.

Applicants should demonstrate why their proposed work could make a meaningful contribution to reducing catastrophic AI risk.

2. Depth

The program is interested in foundational questions that may require substantial time and sustained investigation.

Applicants should therefore think beyond projects that can be completed within a few weeks and instead consider research questions that could generate important insights over a longer period.

Priority AI Safety Research Areas

Although applicants are not restricted to a predetermined list of topics, MATS identifies several areas where it is particularly interested in supporting research.

Safe Automation of Alignment Research

One area of interest involves determining how AI systems can be used to accelerate alignment research while preserving meaningful human oversight.

As AI systems become increasingly capable, researchers face an important challenge: how can AI-assisted research remain checkable, reliable and subject to appropriate oversight?

Research in this area may explore methods for ensuring that AI systems assisting with alignment research do not make the research process impossible for humans to adequately evaluate.

Safety Cases, Evaluations, AI Strategy and AI Assurance

MATS is also interested in work involving:

• AI safety cases.
• AI evaluations.
• AI strategy.
• AI assurance.
• Technical AI governance.

These areas focus on building stronger evidence, methodologies and frameworks that can help organizations and policymakers make informed decisions about advanced AI.

Research may contribute to ways of demonstrating that advanced AI systems meet particular safety requirements or establishing processes for evaluating their risks.

Agentic and Multi-Agent AI Safety

Another priority area concerns agentic and multi-agent AI systems.

As AI systems become capable of acting more autonomously and interacting with other AI agents, new safety and alignment challenges can emerge.

Research in this area may examine how autonomous systems should be designed, evaluated, monitored and aligned when they can independently pursue goals or interact with other systems.

Applicants Are Not Limited to the Priority Areas

The priority areas listed by MATS are intended to provide transparency about the research directions the program is particularly interested in supporting.

They are not intended to prevent applicants from proposing other important AI alignment or AI safety research.

Applicants should propose the work they genuinely believe is important and explain:

• What problem they are trying to solve.
• Why the problem matters.
• What they intend to investigate.
• How their proposed work could contribute to AI safety.
• What evidence would demonstrate progress.
• How the work could ultimately reduce catastrophic risks from AI.

A strong theory of change can therefore be an important component of a competitive application.

MATS Residency Office Locations

Applicants can indicate a preferred office location when applying.

The three locations are:

• London
• Berkeley
• Washington, DC

The Residency follows a hybrid model.

Residents are expected to spend at least one-third of their Residency working in person from the program’s offices.

There is no requirement to remain permanently assigned to one particular office. Residents may spend their in-person time at different MATS offices and can move between locations during the program.

Applicants should therefore select the location that provides the strongest proximity to:

• Relevant AI research laboratories.
• Potential collaborators.
• Researchers.
• Technical communities.
• Policy communities.
• Other resources relevant to their proposed research.

If multiple locations would be equally suitable, applicants can indicate this in their application.

International Applicants

The MATS Residency is open to applicants internationally.

This is particularly important for researchers and professionals based outside the United States and United Kingdom.

For successful applicants who need to relocate internationally, MATS states that it provides visa support.

International applicants should nevertheless carefully consider their work-authorization circumstances and provide the required information in the application.

MATS Residency Application Requirements

Applicants must submit several components through the online application system.

Preparation is therefore important, particularly for applicants who need to contact referees or prepare a representative research sample.

1. Short Profile and Eligibility Section

Applicants begin by providing basic information for application processing and logistics.

This section includes information such as:

• Contact details.
• Current affiliation.
• CV.
• Preferred office.
• Availability for the intake.
• Work-authorization status.
• Highest degree.
• Previous MATS involvement.

The CV must be no longer than two pages.

According to MATS, this section is primarily used for routing and logistical purposes rather than being the central component of the assessment.

2. Track Choice

Applicants must select the MATS track that best fits their circumstances and proposed work.

They are also asked to provide a brief explanation of approximately two to three sentences explaining why the selected track is appropriate.

The track-selection explanation can help reviewers determine whether an applicant may be better suited to another pathway.

3. References

Applicants must provide two references who can directly speak to the quality of their work.

This is especially relevant for applicants whose strongest work is not publicly available.

For example, if an applicant has worked on confidential research or contributed to a larger team, their referee may be able to verify the applicant’s contribution.

Some tracks have additional requirements concerning referees, so applicants should carefully review the requirements for their chosen track.

4. Representative Work Sample

Applicants must submit one work sample that best demonstrates their strongest or most relevant work.

Possible examples include:

• Academic papers.
• Research reports.
• Technical reports.
• Software repositories.
• Evaluations.
• Policy analyses.
• Other substantial research outputs.

The strongest submission is not necessarily the most prestigious document. Applicants should select the work that best demonstrates their ability to conduct rigorous, valuable and independent work.

5. Written Responses

Applicants must complete:

• Two short written responses.
• A brief counterfactual question concerning what they would do if they were not offered a place in the Residency.

Applicants should use these responses to demonstrate their reasoning, motivation and ability to think clearly about their proposed work and future trajectory.

MATS Residency Tracks

The Winter 2027 Residency includes several tracks, each intended for different applicant profiles and career stages.

Bridge Track

The Bridge track is particularly relevant to applicants who may be entering AI safety from another field or who need additional time and support to develop their research portfolio.

Applicants for Bridge must provide a research plan explaining:

• What they intend to work on.
• Why the research matters.
• How it could contribute to AI safety.
• Their theory of change.

Why Bridge?

Applicants should explain why the Residency is particularly valuable at their current stage.

This could involve:

• Transitioning into AI safety.
• Developing a research direction.
• Strengthening an existing body of work.
• Getting dedicated research time.
• Addressing hiring constraints.
• Addressing visa or timing constraints.

Applicants should also explain where they hope to go after the Residency and how participation would help them reach that next stage.

Research Experience and Future Contribution

Applicants should explain how their previous experience prepares them to contribute to AI safety research.

For people transitioning from another field, this is an opportunity to highlight transferable:

• Research methods.
• Technical skills.
• Analytical abilities.
• Policy expertise.
• Domain knowledge.
• Unique perspectives.

Applicants who already have AI safety experience should explain how additional time and support would allow them to deepen their research and prepare for their next opportunity.

Anchor Track

The Anchor track is designed for applicants for whom a longer-term research pathway is the appropriate next step.

Applicants must submit a research plan and provide two referees who have directly supervised or evaluated their work.

Suitable referees may include:

• Research supervisors.
• Collaborators.
• Senior colleagues.
• Other people with direct knowledge of the applicant’s work.

Why Anchor?

Applicants should explain why a two-year staff researcher pathway is the appropriate next step for them.

They should distinguish this pathway from alternatives such as:

• A PhD.
• A postdoctoral position.
• Another research role.
• Other professional opportunities.

Evidence of Independent Work

Applicants must demonstrate that they can take ownership of an open-ended problem and make meaningful progress.

Evidence may include:

• Independent research.
• A thesis.
• Industry research.
• Technical projects.
• Policy work.
• Other comparable experience.

Importantly, applicants do not need peer-reviewed publications or previous AI safety experience to demonstrate this capability.

Incubation Track

The Incubation track is intended for applicants seeking to develop a research agenda, research program or organization.

Instead of submitting a conventional research plan, applicants provide a project plan explaining the direction they intend to develop and its theory of change.

Incubation Rationale

Applicants should explain:

• What is novel about their proposed direction.
• Why the idea requires dedicated time and support.
• Why MATS Incubation is an appropriate environment.
• What they hope to develop during the Residency.

90-Day Plan

Applicants must outline what they intend to test or accomplish during their first 90 days.

A strong plan should also identify what evidence would cause the applicant to:

• Revise the project.
• Pause the project.
• Stop the project altogether.

This demonstrates that the applicant is prepared to learn from evidence rather than simply pursuing a predetermined idea.

Governance and Publication Plan

Where relevant, applicants should discuss:

• Governance.
• Publication strategy.
• Access to research outputs.
• Information hazards.
• Dual-use risks.

This is particularly important for projects involving potentially sensitive AI safety research or technologies that could have both beneficial and harmful applications.

Team and Resources

Applicants should identify the expertise and resources required to execute their project.

For projects involving the creation of an organization, applicants should also identify:

• Roles that still need to be filled.
• Skills required from future team members.
• How they would recruit suitable people.

Use of AI Tools in the Application

MATS allows applicants to use AI tools while preparing their applications.

However, applicants are expected to be transparent about their use of AI.

Applicants should:

1. Briefly disclose whether they used AI tools.
2. Explain how they used them.
3. Confirm that they agree to the applicable AI-use policy.
4. Ensure that the application reflects their own reasoning and judgment.

MATS states that disclosed AI use is not, by itself, a reason for rejection.

The important consideration is whether the application genuinely represents the applicant’s own thinking and work.

This can be particularly useful for applicants whose first language is not English. They may use AI tools to improve the clarity of their writing while retaining responsibility for the substance and reasoning of the application.

Tips for Preparing a Strong MATS Residency Application

Because the Residency emphasizes the quality of work rather than conventional credentials, applicants should focus on demonstrating ability, judgment and potential impact.

1. Develop a Clear Research Idea

Avoid submitting a vague statement about wanting to “work in AI safety.”

Instead, identify a specific problem and explain:

• What the problem is.
• Why it matters.
• What you intend to investigate.
• What existing approaches fail to address.
• What your proposed approach adds.
• What results would change your mind.

2. Focus on the Theory of Change

Your application should make a clear connection between your research and improved AI safety.

Explain how your research could move from:

Research question → findings → practical application → improved safety → reduced catastrophic AI risk.

3. Select Your Strongest Work Sample

Choose the work that best demonstrates your actual capabilities.

A technically sophisticated project, rigorous analysis or strong independent research output may be more useful than simply submitting something because it has a prestigious affiliation.

4. Choose References Carefully

Select people who can provide specific evidence about your abilities.

A referee who has directly observed your research, technical work or analytical skills can potentially provide a much stronger assessment than someone who only knows you superficially.

5. Keep the CV Focused

Because the CV is limited to two pages, applicants should prioritize:

• Relevant research.
• Significant projects.
• Technical or analytical achievements.
• Publications or substantial outputs.
• Relevant professional experience.
• Evidence of independent work.

6. Demonstrate Intellectual Independence

The Residency is interested in people who can tackle difficult, open-ended problems.

Where possible, show examples of situations in which you:

• Identified an important problem.
• Developed your own approach.
• Worked through uncertainty.
• Produced a substantive result.
• Changed your view based on evidence.

7. Do Not Overemphasize Credentials

A lack of a PhD or previous AI safety job should not automatically discourage an applicant.

Instead, use the application to demonstrate what you can actually contribute.

8. Contact MATS if You Are Unsure

Applicants who are uncertain about:

• Their eligibility.
• The most appropriate track.
• The appropriate office.
• The suitability of their research idea.

can contact the MATS Residency team before applying.

The official contact is residency@matsprogram.org.

Why the MATS Residency Winter 2027 Is Worth Considering

The program offers an opportunity for talented individuals to dedicate substantial time to difficult AI safety questions while engaging with an international research community.

Potential advantages include:

• Developing advanced AI safety research experience.
• Working on ambitious and potentially high-impact research questions.
• Connecting with researchers and collaborators.
• Accessing an established AI safety research environment.
• Developing a stronger research portfolio.
• Transitioning from another field into AI safety.
• Pursuing longer-term research through the Anchor pathway.
• Developing a new research project or organization through Incubation.
• Accessing international offices in London, Berkeley and Washington, DC.
• Receiving visa support where international relocation is required.

The precise experience will depend on the applicant’s selected track, research direction and individual circumstances.

How to Apply for the MATS Residency Winter 2027

Interested applicants should begin by reviewing the official Residency information and application requirements.

The application can be submitted online through the official MATS application system.

MATS Residency Winter 2027 Application

Applicants should make sure they have sufficient time to prepare their research or project plan, identify appropriate referees, select a representative work sample and complete the written responses.

APPLY HERE

Important Dates at a Glance

Eligibility and Who Can Apply

Eligible applicant category: International applicants with demonstrated potential to conduct outstanding AI safety research or develop valuable safety-relevant projects.

MATS does not restrict the Residency to people with a particular race, nationality or continent. The program explicitly states that the Residency is open internationally, meaning applicants from Africa, Asia, Europe, North America, South America, Oceania and other regions worldwide may apply, subject to the program’s application and logistical requirements.

Applicants can come from a wide range of professional and academic backgrounds. A PhD, researcher job title, extensive publication record or previous AI safety experience is not required.

Relevant applicants may include:

• Researchers and academics.
• Technical professionals.
• AI and machine-learning specialists.
• Policy researchers and analysts.
• Governance and assurance specialists.
• People transitioning into AI safety.
• Independent researchers.
• Professionals with substantial technical or policy work.
• Entrepreneurs or project founders applying through Incubation.
• Individuals whose strongest work is confidential or unpublished.

Applicants should be able to demonstrate strong reasoning, research ability, seriousness of purpose and the potential to contribute to reducing catastrophic risks from advanced AI.

Visit HERE to learn more about the MATS Residency Winter 2027.

Final Takeaway

The MATS Residency Winter 2027 represents a significant opportunity for people who want to work on some of the most challenging questions surrounding advanced artificial intelligence and its potential risks.

One of the program’s distinguishing features is that it does not rely exclusively on traditional academic credentials. Applicants are assessed primarily on the quality and seriousness of their work and their potential to produce meaningful AI safety research. This creates an opportunity for talented people from academia, industry, policy, technical fields and other backgrounds — including those seeking to transition into AI safety.

The Residency is particularly interested in research involving safe automation of alignment research, AI evaluations and assurance, safety cases, AI strategy, technical AI governance, and agentic and multi-agent AI safety. However, applicants are encouraged to propose other ambitious research directions if they can clearly explain why the work matters and how it could contribute to reducing catastrophic AI risk.

For those considering applying, preparation should begin early. A strong application should contain a focused research or project plan, compelling evidence of previous work, strong references and clear written responses demonstrating independent thinking and sound judgment.

The Winter 2027 application deadline is 31 October 2026, Anywhere on Earth (AoE). Interested applicants should review the official requirements carefully and submit their applications before the deadline.

For more global Fellowship opportunity visit OFY HERE