MATS Winter 2027 streams

Snapshot

Retrieved 2026-08-20 from the official Winter 2027 program page and its linked MATS stream, track, and mentor pages. The published list contains 48 streams across seven tracks. Entries below condense the official copy; follow each stream link for complete wording and updates. Missing fields were not inferred.

Program at a glance

• Main program: 12 weeks, January 19–April 10, 2027.
• Locations: Berkeley, California, and London, UK; remote options are available, and part-time participation may be possible depending on circumstances and mentor.
• Commitment: normally 40 hours/week.
• Application window: August 18–September 6, 2026 (deadline: end of day, Anywhere on Earth).
• Selection: evaluations run through September, October, and early November; offers are expected in early to mid-November.
• Support: $1,600/week\ast \ast forfull-timeparticipation(\ast \ast$19,200 over 12 weeks), plus travel to/from Berkeley or London, housing, office space, and weekday lunch and dinner. Part-time/shorter participation is prorated.
• Extension: selected fellows may continue into a funded 6- or 12-month extension.
• How matching works: Stage 1 applications are to broad tracks. Advancing applicants then choose streams; final matching uses both applicant and stream rankings.

Sources: Winter 2027 program · MATS application · Program FAQ

Track index

Counts include cross-listed streams, so they sum to more than 48.

Track	Published streams
Empirical	22
Biosecurity	11
Policy and Governance	10
Theory	5
Strategy and Forecasting	5
Founding and Field-Building	4
Systems Security	3

Streams

Entries are alphabetical by published stream name.

Abram Demski

• Track: Theory
• Focus: Agent Foundations research focused on clarifying conditions under which humans can justifiably trust artificial intelligence systems. When should one boundedly rational learning-theoretic process come to trust another?
• Mentorship: We can discuss this more and decide on a different structure, but by default, 1 hour 1-on-1 meetings with each scholar once a week, plus a 2 hour group meeting which may also include outside collaborators.
• Good fit: Ability to read and write mathematical proofs; Some fluency with probability theory and expected utility theory.
• Project matching: There will be some flexibility about what specific projects scholars will pursue. Abram will discuss the current state of his research with scholars and what topics scholars are interested in, aiming to settle on a topic by or before week 2.
• Sources: stream details · mentor profile

Active Site

• Track: Biosecurity
• Focus: In my stream, I focus on empirically measuring how frontier AI changes human capabilities in biology and the implications of it for biological risk. I’m particularly interested in real-world evaluations, understanding what drives AI-enabled uplift, and developing simulations or other proxies that can be validated against physical experiments and used to rapidly assess new models.
• Good fit: Research Independence. Has previously taken ownership of a research project and can make progress without frequent direction on what to do next; Research Focus. Ability to narrowly scope a larger question into a project that can produce a useful result during the program; Technical Skills. Has sufficient technical skill to perform the core work of the project;…
• Sources: stream details · mentor profile 1 · mentor profile 2

Aidan O’Gara

• Track: Policy and Governance
• Focus: This stream will focus on compute governance, including data center inspections, export controls, and distributed training.
• Sources: stream details · mentor profile

AI Futures Project

• Track: Strategy and Forecasting
• Focus: We are interested in mentoring projects in AI forecasting and governance. This work would build on the AI 2027 report to either do more scenario forecasting or explore how to positively affect key decision points, informed by our scenario.
• Mentorship: We will have meetings each week to check in and discuss next steps. We will be consistently available on Slack in between meetings to discuss your research, project TODOs, etc.
• Good fit: Strong reasoning and writing abilities; Excitement about AI forecasting/governance research; Autonomous research skills; Ability to learn quickly; Significant background knowledge in at least one area relevant to AI forecasting and governance (e.g. government experience, technical background, etc.); Significant background knowledge in AGI and existential risks
• Project matching: We will talk through project ideas with scholar
• Sources: stream details · mentor profile 1 · mentor profile 2 · mentor profile 3 · mentor profile 4

Alfie Lamerton and Joe Kwon

• Track: Empirical
• Focus: This stream focuses on secret loyalties, where an LLM covertly tries to advance a principal’s interests. Secret loyalties have been established as a pressing threat [1], and model organisms of narrow secret loyalties have been constructed and audited [2]. This stream aims to advance the empirical foundations of our understanding of secret loyalties.…
• Good fit: ML research and engineering experience. This will be a substantial process taking up a large portion of mentees’ time. Doing well at this will enable generation of results and iteration which are important to a project like this. We are talking about working in git repositories, handling experiments, tracking and diagnosing results effectively, generating hypotheses and research questions,…
• Sources: stream details · mentor profile 1 · mentor profile 2 · 1(https://www.formationresearch.com/secret-loyalties-whitepaper.pdf) · 2(https://arxiv.org/abs/2605.06846)

Alignment Research Center (ARC)

• Track: Theory
• Focus: The Alignment Research Center is a small non-profit research group based in Berkeley, California, that is working on a systematic and theoretically grounded approach to mechanistically explaining neural network behavior. We are interested in fellows with a strong math background and mathematical maturity. If you’d be excited to work on the research direction described in this blog post – then we’d encourage you to apply!
• Mentorship: Scholars will work out of ARC’s offices in Berkeley. Each scholar will meet with their mentor at least once a week for an hour, though 2-3 hours per week is not uncommon. Besides time with their official mentor, scholars will likely spend time working in collaboration with other researchers;…
• Good fit: Mathematical maturity and a math, physics or computer science background at the level of a strong undergraduate at a top-20 university; Good at communicating about technical topics; Interest in engaging with ARC’s higher-level research agenda; Potentially interested in joining ARC full-time by January 2028.
• Project matching: Each scholar will be paired with the mentor that best suits their skills and interests. The mentor will discuss potential projects with the scholar, and they will decide what project makes the most sense, based on ARC’s research goals and the scholar’s preferences.…
• Sources: stream details · mentor profile 1 · mentor profile 2 · mentor profile 3 · mentor profile 4 · mentor profile 5 · this blog post

Aman Patel, Adin Richards

• Track: Biosecurity
• Focus: This stream is primarily focused on research into physical defenses against engineered pathogens, aiming to inform decisions about PPE stockpiling and distribution approaches, improve improvised PPE and bioshelter scale-up, and reach rapid conclusions on how much to prioritize other areas of physical biodefense (agriculture, emergency response, etc.). We are also open to strategic research into the use of bioweapons by AI or AI-human teams as part of takeover strategies and how this might inform preparedness.
• Mentorship: One 30-60 min weekly meeting by default. We’re active on slack and can usually respond to quick questions there within the work day. For more substantive async engagement, especially project feedback, google doc comments are probably best.
• Good fit: Familiar with developing BOTECs; Willing to quickly get up to speed in new technical areas; Capable of working independently; Happy to pivot in response to findings or feedback
• Project matching: We’ll provide fellows with a short list of projects and will meet with them to discuss which they feel most excited about / best suited for.
• Sources: stream details · mentor profile 1 · mentor profile 2

Anthropic

• Track: Empirical
• Focus: This coalition of mentors make up the “Anthropic Stream”. This stream spans a range of empirical research areas in AI safety on LLMs, including AI control, scalable oversight, model organisms, model internals, model welfare, security, and more. You’ll be pitched, and have the option to pitch, a variety of safety research projects, and then be matched to projects and mentors based on your interests/preferences on research and what you’d like to get out of MATS.…
• Mentorship: During the program, scholars meet weekly with their project mentors and collaborators. Some projects meet more often without mentors (e.g., daily standups with the peers on the project). Each project will have a primary mentor, who is also the main decision-maker on key milestones for the project and who is the default person to go to for feedback, advice,…
• Good fit: See the top of this post Generally someone who can run a lot of experiments quickly.
• Project matching: Mentorship starts with the “Project Pitch Session” Anthropic runs at the start of the program. Fellows get ~1 week to derisk and trial projects before submitting their preferences. Starting on week 2, scholars are assigned projects where the primary mentor is whoever pitched it.…
• Sources: stream details · mentor profile 1 · mentor profile 2 · mentor profile 3 · mentor profile 4 · mentor profile 5 · this post

Apollo Research - Monitors

• Track: Empirical
• Focus: We will continue working on black-box monitors for scheming in complex agentic settings, building on the success of the previous stream. Concretely, we will work on scaling our datasets and fine-tuning efforts, as described in the scalable monitoring agenda Most likely the next projects will be about automated iterated red-team vs. blue-team games. We are currently training the blue team. We will then train the red-team and within this stream, we will try and close the loop to train them both synchronously.
• Mentorship: We have two weekly 60-minute calls by default. Since everyone will work on the same project, these calls will be with all participants of the stream. I respond on slack on a daily basis for asynchronous messages. Scholars will have a lot of freedom for day-to-day decisions and direction setting. In the best case,…
• Good fit: You like quick empirical iteration and direct feedback loops. I think candidates who did well in the past were good at producing a high volume of output, running many experiments in parallel and keeping track of what the most important next steps are at any given point in time; You enjoy tinkering with LLMs, e.g. prompting, building basic LM agents, and fine-tuning;…
• Project matching: You will work on subprojects of black box monitoring. See here for details.
• Sources: stream details · mentor profile · scalable monitoring agenda · here

Caspar Oesterheld (Redwood, conceptual reasoning capabilities)

• Track: Empirical · Theory
• Focus: Theory of change: Soon, most important work will be done by AI. AI is going to increasingly advise people and help with important things, many of which are time-sensitive and path dependent, e.g., work on alignment/safety (including various things like how LLMs should behave given that they’re very persuasive); how to think about acausal trade; how to organize society. It seems good for AI to do well at those things. Of course,…
• Good fit: None
• Sources: stream details · mentor profile · conceptualreasoning.ai

Damon Binder

• Track: Biosecurity · Strategy and Forecasting
• Focus: This stream focuses on how advanced AI could enable new and dangerous bio technologies, and on assessing when risks become tractable or urgent as those capabilities arrive.
• Mentorship: Half-hour one-on-one weekly meetings by default, with the option to extend or add ad-hoc calls when useful. I’m active on Slack and typically respond within a day for quick questions. I’m happy to read drafts and leave written feedback async between meetings.
• Good fit: Eagerness to use and experiment with AI tools in novel ways; Research independence; Intellectual breadth and curiosity. Someone excited to work across many domains of science and to learn a lot in the process; Strong quantitative background; Background in a natural science or engineering; Prior experience driving a research project to completion; Some coding experience, ideally in Python;…
• Project matching: I’ll talk with the fellow about what they’re interested in, and we’ll pick a broad area together from a few directions I’d want to pitch. From there we’ll work together to scope something sharp and well-defined, with me leaning on my sense of what’s tractable and high-value. The fellow then runs with the project,…
• Sources: stream details · mentor profile

Dave Banerjee

• Track: Policy and Governance · Strategy and Forecasting
• Focus: My stream focuses on preserving checks and balances as governments adopt increasingly powerful AI. Fellows will work on questions like how Congress can maintain oversight of an AI-accelerated executive branch (including via privacy-preserving AI auditors) and what a positive vision for government AI adoption looks like. Projects will typically produce a public report and sometimes involve engaging directly with policymakers and other stakeholders.
• Good fit: Strong conceptual reasoning; Strong writer (or at least has the potential to become one); Highly motivated to learn and self-improve; Good organizational skills. Doesn’t drop balls.
• Sources: stream details · mentor profile 1 · mentor profile 2

David Africa

• Track: Empirical
• Focus: This stream will focus on model motivations and character, open-ended environments, and new forms of misalignment.
• Good fit: Receptive to feedback! --- It’s bad to be conflict averse and not tell me when things are going wrong, or say that you are gonna do a thing and then not do it. It’s fine and often good to fail fast; Fast / agentic --- it’s important that you use your best judgement and often! Don’t be paralyzed; Knows when and where to be cyborged up --- the optimal amount of slop, esp to execute fast, is not zero.…
• Sources: stream details · mentor profile

David Lindner

• Track: Empirical
• Focus: This stream will focus on monitoring, stress-testing safety methods, and evals, with a focus on risks from scheming AIs. Examples include (black-box) AI control techniques, white-box monitors (probes etc.), chain-of-thought monitoring/faithfulness, building evaluation environments, and stress-testing mitigations.
• Mentorship: For each project, we will have a weekly meeting to discuss the overall project direction and prioritize next steps for the upcoming week. On a day-to-day basis, you will discuss experiments and write code with other mentees on the project (though I’m available on Slack for quick feedback between meetings or to address things that are blocking you).…
• Good fit: I am looking for scholars with strong machine learning engineering skills, as well as a background in technical research. While I’ll provide weekly guidance on research, I expect scholars to be able to run experiments and decide on low-level details fairly independently most of the time. I’ll propose concrete projects to choose from, so you should not expect to work on your own research idea during MATS.…
• Project matching: We will most likely have a joint project selection phase, where we present a list of projects (with the option for scholars to iterate on them). Afterward, each project will have at least one main mentor, but we might also co-mentor some projects.
• Sources: stream details · mentor profile

Dewi Erwan (BlueDot Impact)

• Track: Founding and Field-Building
• Focus: Founding ambitious AI safety and field-building projects.
• Mentorship: Minimum support = 2x 30-min meetings per week. We could scale this up as appropriate. I’ll be based in SF. If the fellows want to work in BlueDot’s office for some periods of time, I could collaborate with them daily. I’m available for quick calls anytime, and am responsive on Slack.
• Good fit: I’m open to people with a wide range of backgrounds. Though you need to be willing to work very hard, be great at communicating, and have a burning desire to make AI go well. I work best with people who are intense, communicate and reason clearly, and are mission-driven.
• Project matching: We’ll work together to design the project. You’ll have a lot of freedom to figure out what the best shape of thing to do is, and I’ll provide lots of regular feedback and make relevant introductions to help you refine the proposal. Your first 1-2 weeks will be focused on figuring out what to do,…
• Sources: stream details · mentor profile

Fourth Eon Biosecurity

• Track: Biosecurity · Empirical
• Focus: At Fourth Eon Biosecurity we’re building adaptive, AI-native safeguards across the bioengineering stack, with a focus on function-based DNA synthesis screening. Fellows in this stream will work on technical research projects at the intersection of AI safety and biosecurity, aimed at reinforcing screening and generalizing detection beyond known threat signatures. Projects span mechanistic interpretability of bio foundation models, model evaluations for biosecurity-relevant capabilities,…
• Mentorship: I typically schedule a standing weekly 1:1 meeting with each fellow, and also hold a weekly research group meeting. Beyond that I am available on Slack and can find additional time for calls outside of scheduled meetings. Note that as part of our Safe and Responsible Research Framework we require fellows to sign a fellowship agreement covering…
• Good fit: Prior technical research experience; Strong critical thinking and creative problem-solving abilities; Curiosity and a desire to understand the world; The integrity and judgment to responsibly carry out sensitive research; Expertise in one or more of the following domains: bioinformatics, computational biology, structural biology, biochemistry, molecular biophysics, protein engineering, biosecurity, AI/ML,…
• Project matching: Fellows who are interested in our research area should think of potential project ideas that leverage their strengths and interests. I will work with individual fellows to identify a specific project that matches their background and interests and is aligned with our overall research direction,…
• Sources: stream details · mentor profile

Gabriel Kulp

• Track: Systems Security
• Focus: In this project, we will explore GPU side-channel attacks to extract information about model usage. A simple example is to observe (via radio, power fluctuations, acoustics, etc.) which experts were used in each forward pass of an MOE model, then use those observations to guess which tokens were produced.
• Mentorship: Co-working 2-4 hours per week, including detailed guidance. Flexible. 1 hour check-ins per week. You can schedule ad-hoc calls if stuck or wanting to brainstorm.
• Good fit: Machine learning: familiarity with LLM architecture, especially for mixture-of-experts. Making classifier, regression, and generative models for unusual data types; Math: statistical tests of correlation and mutual information; Electrical engineering: signal processing, transmission lines, antennae and radio, switching power supplies; Hardware: GPU architecture (memory and compute), PCIe traffic,…
• Project matching: There is a cluster of potential projects to choose from. As a team, we will decide which to pursue based on individual interest and skills. Mentors will pitch example projects and scholars can then modify and re-pitch them. Once the research problem, hypothesis, and testing plan are written and agreed on,…
• Sources: stream details · mentor profile

Jacob Swett

• Track: Biosecurity
• Focus: This stream will focus on projects related to biosecurity countermeasures.
• Mentorship: 1 hour weekly meetings by default for high-level guidance. Onboarding to our slack, which has access to the entire Blueprint Biosecurity team. Can be reached async every day and can meet as needed.
• Good fit: Have some sort of a technical background (e.g. technical undergraduate degree); Have familiarity with reading research papers; Are agentic; Prioritize truthseeking; Are focused on maximizing impact
• Project matching: Far-UVC; Germicidal UV (GUV); Glycol Vapors; Disinfectant Vapors; In-Room Filtration; PPE; Nasal Sprays; Diagnostics; Biosurveillance
• Sources: stream details · mentor profile

Joshua Engels

• Track: Empirical
• Focus: I’m interested in better understanding and controlling how post-training causes alignment-relevant behavior. This is a pretty broad area, and I’m open to many approaches to these problems! Potential areas of study / methods of attack might include model organisms, training run science/ablations, root causing strange behaviors, or studying how best to robustly induce behaviors or values or beliefs into models.
• Good fit: Research experience: at least one project where you drove the research direction (first-author or equal contribution); Fast iterators: You are good at derisking: quickly figuring out whether projects will work or not. You are calibrated on your work and can e.g. figure out whether to dig into a result you are skeptical of or move on to a new experiment; Familiarity with machine learning and LLMs
• Sources: stream details · mentor profile

Juniper Ventures

• Track: Founding and Field-Building
• Focus: This stream will focus on biosecurity, behavioral science for AI governance, and evaluating interventions to reduce catastrophic risk.
• Sources: stream details · mentor profile 1 · mentor profile 2

LILA (David Peinador Veiga)

• Track: Biosecurity · Empirical
• Focus: The stream focuses on evaluating and/or mitigating catastrophic risk emerging from dangerous scientific capabilities in frontier AI systems, with an emphasis on the challenges that emerge from lab integrations and novel science. Potential research directions include evaluation design, risk mitigations and evaluation science.
• Mentorship: We can schedule a weekly 1h meeting, for general progress updates, sharing results, and overall guidance. I would be reachable on Slack as well for async comms. Happy to jump on ad-hoc calls for specific discussions or pair coding/debugging. I am based in London and I work UK hours (10am-7pm), but I also visit the US (Boston) a few times a year.
• Good fit: One of two backgrounds: either (a) hands-on experience building or running LLM evaluations, or (b) applying AI/ML to a scientific domain (bio, chem, materials); Research experience: at least one project where you drove the direction rather than executing someone else’s plan. Academic and industry routes are both fine;…
• Project matching: I will work with the fellow to find the right project that suits their interest within the directions spelled out above. I will pitch a few project ideas and support the fellow in making the decision. I also welcome project suggestions; in those cases I would work with the fellow to scope it appropriately.
• Sources: stream details · mentor profile

Lisa Thiergart, Luis Cosio, Guy (SL5 Task Force)

• Track: Systems Security
• Focus: The SL5 Task Force will build out a prototype SL5 datacenter this year together with frontier AI labs. This will be a massive research and engineering project with many avenues for spinning out new organizations and research programs. This project is urgent due to this technology being needed in the next 1 to 2 years.
• Mentorship: We will meet at least 1h a week synchronously and communicate daily via stand-ups on slack. I typically respond within a few hours for additional feedback and within 1-3 days for indepth code or other review. Scholars can also schedule adhoc calls with me or my co-mentor Luis if they’re stuck.…
• Good fit: BSc Computer Science or Electrical engineering, or comparable experience; 3+ years of Security Engineering, Infrastructure Engineering or related experience - demonstrated ability to independently lead complex projects as an IC relating to designing, operationalising and implementing security programs/controls;…
• Sources: stream details · mentor profile 1 · mentor profile 2 · mentor profile 3

Luke Drago, Rudolf Laine

• Track: Strategy and Forecasting · Policy and Governance
• Focus: This stream focuses on identifying tractable policy and technical interventions to gradual disempowerment, focusing on economic disempowerment and the intelligence curse. Possible project areas include: Formalizing intelligence curse dynamics into a model that can be tracked and monitored. More durable policy solutions to mass unemployment than UBI. Technical interventions to extend the centaur period.
• Mentorship: We’ll meet 1:1 for 30 minute slots twice a week, once with each mentor. We’ll be active on Slack (default to over-slacking us), and can do quick ad-hoc calls as well. Once a week, we expect you to have some artifact that we will give feedback on.
• Good fit: Good at thinking across disciplines. Fellows will need to think across machine learning, economics, political theory, and historical precedents; Generativeness. We want people who can come up with new things that don’t fit within existing ideological or intellectual frameworks;…
• Project matching: We provide three projects as options we are excited about, but they are not inclusive of all ideas. During the application process, we will ask potential mentees to either a) sharpen these proposals into a more specific question incorporating their own interests, or b) propose their own projects.…
• Sources: stream details · mentor profile 1 · mentor profile 2

Maksym Andriushchenko

• Track: Empirical
• Focus: This stream focuses on critical challenges in AI safety and alignment, including risks from automating AI research, bottlenecks to recursive self-improvement, and the automation of safety and alignment research. Priority topics also include AGI privacy, measuring long-horizon agentic capabilities, developing new alignment methods, and advancing the science of post-training.
• Mentorship: I usually spend at least 30 min per week in one-on-one meetings with my mentees. We can also discuss longer time slots if necessary. Besides these time slots, I try to be as responsive as possible over Slack (>2 comprehensive responses per day) and read relevant papers between weekly meetings.
• Good fit: Prior research experience in a topic related to AI safety (at least one completed project with first-author contribution); Independent, self-driven personality; Strong general computer science background; Ideally, a good software engineering background; Familiarity with deep learning frameworks; Clear communication
• Project matching: I would prefer to set the overall direction, but I will listen closely to scholars about their preferences within a broad direction. Converging on a particular topic is expected to be a collaborative process.
• Sources: stream details · mentor profile

Matthew Gentzel

• Track: Policy and Governance
• Focus: Escalation risks from state perceptions of AI capability, AI-enabled targeting, AI-enabled decision manipulation, and the impact of AI integration into nuclear command and control.
• Mentorship: Mentorship will mostly consist of calls, sorting through research ideas and providing feedback. I’ll be up to review papers, and potentially to meet in person depending on timing.
• Good fit: Looking for intellectually curious and honest scholars, with some background on topics related to national security, game theory, or AI-enabled military and influence capabilities.
• Project matching: I’ll talk through project ideas with scholar, or the scholar can pick from a list of projects
• Sources: stream details · mentor profile

Mauricio Baker, Anjay Friedman

• Track: Systems Security · Policy and Governance
• Focus: This stream focuses on AI policy, especially technical governance topics. Tentative project options include: technical projects for verifying AI treaties, metascience for AI safety and governance, and proposals for tracking AI-caused job loss. Scholars can also propose their own projects.
• Mentorship: We’ll meet once or twice a week (~1 hr/wk total, as a team if it’s a team project). I’m based in DC, so we’ll meet remotely. I (Mauricio) will also be available for async discussion, career advising, and detailed feedback on research plans and drafts.
• Good fit: No hard requirements. Bonus points for research experience, AI safety and governance knowledge, writing and analytical reasoning skills, and experience relevant to specific projects.
• Project matching: I’ll talk through project ideas with scholar
• Sources: stream details · mentor profile 1 · mentor profile 2

McKenna Fitzgerald

• Track: Policy and Governance
• Focus: This stream will focus on preparing AI governance policies for future policy windows through scenario mapping and policy architecture.
• Good fit: General understanding of US political system incl. legislative process; Understanding of current US AI policy legislation and current conversations; Strong written communications skill (bonus for policy-specific writing); Demonstrated analytical and research ability; Demonstrated intellectual independence and willingness to update
• Sources: stream details · mentor profile

Michael Chen

• Track: Empirical · Policy and Governance
• Focus: Research papers (technical governance or ML) related to evaluating and mitigating dangerous AI capabilities, with a focus on what’s actionable and relevant for AGI companies
• Mentorship: I like to get daily standup messages about progress that has been made on the project, and I’m happy to provide some quick async feedback on new outputs. I’ll also have weekly meetings. I’m based in Constellation in Berkeley.
• Good fit: Good writers/researchers who can work independently and autonomously! I’m looking for scholars who can ship a meaningful research output end-to-end and ideally have prior experience in writing relevant papers.
• Project matching: I may assign a project, have you pick from a list of projects, or talk through project ideas with you.
• Sources: stream details · mentor profile

Neel Nanda

• Track: Empirical
• Focus: Neel takes a pragmatic approach to interpretability: identify what stands between where we are now and where we want to be by AGI, and then focus on the subset of resulting research problems that can be tractably studied on today’s models. This can look like diving deep into the internals of the model, or simpler black box methods like reading and carefully intervening on the chain of thought - whatever is the right tool for the job. This could look like studying how to detect deception,…
• Sources: stream details · mentor profile · this podcast

Oliver Crook

• Track: Biosecurity · Empirical
• Focus: Computational/modelling problems in biosecurity.
• Mentorship: Typically 1 hour weekly meetings by default. I typically respond on slack quite quickly - some weeks I am not available. You are welcome to chat to my phd students too!
• Good fit: Computational experience e.g. Python OR statistical modelling interest in biosecurity
• Project matching: We will construct a project together that best suits the skills and interests of the fellow and what I can reasonably be helpful for.
• Sources: stream details · mentor profile

Patricia Paskov & Miles Brundage (AVERI)

• Track: Policy and Governance
• Focus: This stream will focus on AI auditing, evaluation standards, and governance.
• Good fit: Familiar with the frontier AI auditing/evaluation ecosystem and discourse (e.g. evaluators, major organizations, proposals, policies/legislation, etc.); Strong analytical writing skills; Has synthesized a complex landscape into a structured artifact — a database, taxonomy, ecosystem map, flagship report, dashboard, or paper; Works independently and takes initiative,…
• Sources: stream details · mentor profile 1 · mentor profile 2

Patrick Butlin

• Track: Theory · Empirical
• Focus: Projects in this stream will be on AI welfare and moral status; more specifically, on what it takes to be a moral patient and how we can determine whether AI systems meet the conditions. I’m looking for applicants who have ideas about these topics and are motivated to explore them in more detail.
• Mentorship: By default, scholars will meet with me online for 1hr/week and I will respond to questions on email/slack.
• Good fit: Motivation to work in this area and curiosity about the ideas; Initiative and ability to work independently; Excellent general intellectual and communication skills; Strong background in a relevant area, such as ethics, philosophy of mind, cognitive science (especially topics such as consciousness science or human reinforcement learning), or experiments with LLMs
• Project matching: I will talk through project ideas with scholar
• Sources: stream details · mentor profile

Raymond Douglas & David Duvenaud

• Track: Strategy and Forecasting
• Focus: This stream will explore AI slowdown dynamics, desirable futures, moral convergence, and open-ended forecasting.
• Sources: stream details · mentor profile 1 · mentor profile 2

Redwood Research

• Track: Empirical
• Focus: The Redwood Research stream is looking for fast empirical iterators and strategists to work on control research.
• Mentorship: 30-60 min weekly meeting; potentially daily stand-ups; Slack messages
• Good fit: fast at empirical ML iteration; thoughtful and articulate about AI safety; strong at quantitative reasoning.
• Project matching: We will assign projects by default but are open to getting pitched on projects.
• Sources: stream details · mentor profile 1 · mentor profile 2 · mentor profile 3 · mentor profile 4 · mentor profile 5

Richard Ngo

• Track: Theory
• Focus: My MATS fellows will do philosophical thinking about multi-agent intelligence and how agents change their values. This will likely involve trying to explore and synthesize ideas from game theory, signaling theory, reinforcement learning, and other related domains.
• Mentorship: I’ll come meet scholars in person around 2 days a week on average. On those days I’ll be broadly available for discussions and brainstorming. On other days scholars can message me for guidance (though I’d prefer to spend most of my effort on this during the in-person days).
• Good fit: My main criterion for selecting scholars will be clarity of reasoning.
• Project matching: I will talk through project ideas with the scholar.
• Sources: stream details · mentor profile

Roger Grosse

• Track: Empirical
• Focus: Roger Grosse’s stream investigates how to improve influence functions and other training data attribution methods, and uses these tools to study alignment-related phenomena such as out-of-context reasoning and emergent misalignment. The ideal scholar has experience with LLM internals, strong statistics/applied math skills (especially numerical linear algebra), and can independently drive research from literature review through experimentation and analysis.…
• Mentorship: I will meet with scholars 1 hour per week by default, and will be available to answer questions on Slack roughly daily.
• Good fit: Experience working with LLM model internals; Strong background in statistics and/or applied math (esp. numerical linear algebra); Ability to carry out research independently on the timescale of weeks (reading the literature, formulating and carrying out experiments, interpreting results); Ability and willingness to dig into details to get at the root causes of phenomena
• Project matching: I will give the scholar the level of freedom they are ready for. I will be prepared with focused, shovel-ready projects, but exceptional scholars with a vision they are excited about will have the flexibility to pursue it.
• Sources: stream details · mentor profile

Rosie Campbell (Eleos)

• Track: Founding and Field-Building
• Focus: Eleos AI Research is a nonprofit working to ensure the interests of AI systems are appropriately taken into account as we navigate transformative AI. We are looking for competent generalists to help scale up our AI welfare field-building by running high-impact events, or to otherwise amplify our research, communications, and operational capacity.
• Mentorship: By default, I expect to meet with each fellow for at least 60 minutes per week (possibly split into 2 x 30 minute meetings). For projects that require a lot of input (e.g. event organizing), we can scale that up as needed. I’ll also be available for ad hoc meetings, and can be reached asynchronously on Slack or by email. At Eleos,…
• Good fit: Has experience managing projects or events; Has excellent written and verbal communication skills; Is conscientious and detail-oriented; Can navigate multiple stakeholders with different interests and incentives; Knows when to take initiative and solve problems independently, and when to seek help and input; Is familiar with the field of AI welfare
• Project matching: We’ll meet at the start of the program to discuss ideas in the areas listed above, as well as ideas that fellows would like to pitch. We’ll jointly decide on a project that aligns with Eleos’s priorities as well as the goals and skills of the fellow.
• Sources: stream details · mentor profile

Ross Matican & Mike McCormick (Halcyon Futures)

• Track: Founding and Field-Building
• Focus: Backing projects focused on product development and organization building in the areas of AI safety and alignment, biosecurity, and critical cybersecurity. Looking for fellows who are self starters, default to action, and have a desire to create.
• Mentorship: Co-mentorship from Halcyon’s Ross Matican (Investor & Grantmaker), Mike McCormick (Founder, CEO), and Charlie Petty (Venture Partner). Ross will be leading point. Scheduled 45 min bi-weekly meetings (every other week). Ad hoc meetings can be added between scheduled sessions. We’ll have a shared Slack channel with Ross, Mike, and Charlie,…
• Good fit: Strong technical hardskills (e.g., MLE, SWE, math); Domain expertise in AI safety and alignment, biosecurity, or cybersecurity; Experience with building or working on a usable product; Openness to iterating, pivoting, and failure; General understanding of business and organization principles; Considerate and works well in team settings; 5+ years of relevant experience: recruiting, talent / HR, executive search,…
• Project matching: For product development and organization building projects in the areas of AI safety and alignment, biosecurity, and critical cybersecurity - fellows will have full freedom. We expect fellows to come with rough ideas and opinions on direction that will inform where they start exploring the market.…
• Sources: stream details · mentor profile 1 · mentor profile 2

Safe AI Forum

• Track: Policy and Governance
• Focus: We work to advance technically grounded international coordination to reduce catastrophic risks from frontier AI, with a particular focus on China and Western countries. We translate technical AI safety and governance tools into practical proposals for cooperation.
• Mentorship: 1 hour weekly meetings by default for high-level guidance. We are active on Slack and typically respond within a day for quick questions.
• Good fit: A strong working understanding of frontier AI governance and safety, including major international initiatives and institutions such as the AI Safety Institute network, AI Safety Summits, and relevant multilateral processes; Familiarity with Chinese AI governance and the Chinese AI ecosystem, including key institutions, policy developments, safety debates, and relevant technical or regulatory trends;…
• Project matching: We will provide a shortlist of projects that we are keen for the scholar to work on in Week 1. We’ll ask scholars to scope these in the 1st week and make a determination about which project to focus on in Week 2.
• Sources: stream details · mentor profile 1 · mentor profile 2

Sebastian Oehm, Askar Kleefeldt

• Track: Biosecurity
• Focus: Therapeutics may have durable advantages over pathogens even in the limit of technological progress. How can therapeutic development and manufacturing be made resilient under biorisk scenarios? How can AI progress be maximally leveraged for defense?
• Mentorship: I expect we will spend some time at the beginning scoping out a project that is a good fit for the fellow’s background and interests. Then, project supervision will depend strongly on the nature of the project. Generally, I expect the fellow to take ownership of the work,…
• Good fit: Strong interest in biosecurity, AI, and therapeutics. The fellow should be motivated by the question of how societies can maintain and accelerate therapeutic response capacity under severe biological risk scenarios; Research independence. The fellow should be able to take an ambiguous question, break it into tractable parts, identify relevant evidence, and converge on a useful output;…
• Project matching: We will jointly define the exact project with the fellow, based on their background, interests, and comparative strengths, as well as our current priorities. I expect strong fellows may have their own questions and ideas, but we will provide substantial guidance early on to help turn those ideas into a clear, useful,…
• Sources: stream details · mentor profile 1 · mentor profile 2

SecureBio AI

• Track: Biosecurity · Empirical
• Focus: This stream will work on projects that empirically assess national security threats of AI misuse (CBRN terrorism and cyberattacks) and improve dangerous capability evaluations. Threat modeling applicants should have a skeptical mindset, enjoy case study work, and be strong written communicators. Eval applicants should be able and excited to help demonstrate concepts like sandbagging elicitation gaps in an AI misuse context.
• Mentorship: Typically, this would include weekly meetings, detailed comments on drafts, and asynchronous messaging.
• Good fit: Skeptical mindset; Transparent reasoning; Analytical; LLM engineering skills (e.g., agent orchestration); Biosecurity knowledge
• Project matching: Mentor(s) will talk through project ideas with scholar
• Sources: stream details · mentor profile

Sunishchal Dev

• Track: Biosecurity · Empirical
• Focus: This stream will focus on evaluation science, CBRN capability benchmarks, and automated red-teaming.
• Sources: stream details · mentor profile

TAIGR

• Track: Policy and Governance
• Focus: This stream will focus on technical AI governance research — hence the name TAIGR. We will follow an academic collaboration model and produce open research on applied AI safeguards, incidents, laws, and other impactful topics in AI governance.
• Mentorship: By default, we should expect to meet 2-3 times per week as a full group, plus ad hoc project-specific meetings.
• Good fit: Experience in conducting, writing, and presenting academic research.
• Project matching: I will work with MATS scholars to iteratively refine project ideas in whatever area our interests and skills overlap. Above all, project selection will hinge on having a clear (and good) theory of impact.
• Sources: stream details · mentor profile

Team Shard

• Track: Empirical
• Focus: In the shard theory stream, we create qualitatively new methods and fields of inquiry, from steering vectors to gradient routing to unsupervised capability elicitation to robust unlearning. If you’re theory-minded, maybe you’ll help us formalize shard theory itself.
• Mentorship: We will have weekly 1-1’s and weekly team lunch, as well as asynchronous communication over Slack. Mentees are always welcome to reach out at any time, in case guidance is needed outside of usual meeting times. Scholars should mostly figure things out on their own outside of meetings
• Good fit: Academic background in machine learning, computer science, statistics, or a related quantitative field; Familiarity with ML engineering; Proven experience working on machine learning projects, either academically or professionally; Strong programming skills, preferably in Python, and proficiency in data manipulation and analysis; Ability to write up results into a paper.
• Project matching: Mentor(s) will talk through project ideas with scholar
• Sources: stream details · mentor profile 1 · mentor profile 2

Toby Webster

• Track: Biosecurity
• Focus: This stream will focus on technical and policy work at the intersection of AI and biosecurity.
• Sources: stream details · mentor profile

Tomek Korbak

• Track: Empirical
• Focus: We are interested in AI control and scalable oversight. I’m excited to work with scholars interested in empirical projects building and evaluating control measures and oversight techniques for LLM agents, especially those based on chain of thought monitoring. I’m also interested in the science of chain of thought monitorability, misalignment and control. An ideal project ends with a paper submitted to NeurIPS/ICML/ICLR.
• Mentorship: I’ll meet with mentees once a week and will be available on Slack daily.
• Good fit: An ideal mentee has a strong AI research and/or software engineering background. A mentee can be a PhD student and they can work on a paper that will be part of their thesis.
• Project matching: I’ll talk through project ideas with scholar
• Sources: stream details · mentor profile 1 · mentor profile 2

Transluce

• Track: Empirical
• Focus: We build scalable technology for AI understanding and oversight.
• Mentorship: You will work closely with a mentor through recurring meetings (group and individual) and Slack.
• Good fit: We’re looking for strong, experienced software engineers or talented researchers who can hit the ground running and iterate quickly. ML experience is a bonus but not required.
• Project matching: We will talk through project ideas with scholar
• Sources: stream details · mentor profile 1 · mentor profile 2 · mentor profile 3

Truthful AI

• Track: Empirical
• Focus: This stream will focus on evaluating dangerous capabilities in language models and detecting deception and dishonesty.
• Sources: stream details · mentor profile 1 · mentor profile 2

Notes on interpretation

• A track is a broad research area; a stream is one or more mentors working on a shared research agenda.
• Several streams are cross-listed under multiple tracks.
• This is a discovery and comparison note, not a substitute for the full stream pages. Application requirements and project menus can change.