Hide table of contents

Throughout this post I’m speaking in my personal capacity, not as a MATS representative. In particular, I’m drawing from my experience of RM in London in 2025, mostly covering the extension phase of the MATS program. Aspects of this may not be representative of broader research management at MATS or the main program in Berkeley.

About my Job: Research Manager

I’ve been a Research Manager at MATS since the start of 2025. Most weeks, people receiving 80k career advising asked me about what my work actually involves, so I thought I’d write a post with my thoughts on the role.

This post was originally intended to be posted alongside @frances_lorenz’ post on my career transition during Career Conversations Week and inspired by other great posts. However, through my delay in posting it now takes on a different context, a fond reflection of my time at MATS as it draws to an end.

Background

I studied Aerospace Engineering at Bristol and after not securing a role in the UK’s modest space sector, I went to London to work in tech roles in financial services. For ten years I was a technical product manager, building benchmarking and analytics platforms in capital markets.

How I got here

Frances recently wrote a great piece about my career transition, which you can read here.

What is Research Management?

Scholars at MATS work with brilliant mentors who provide research supervision for their projects. They are also paired with a research manager who provides additional support. My approach to research management is centred around a strong commitment to servant leadership and radical candor.

I am here to help the scholars (and their mentors) get the most out of MATS by finding ways in which we can support them, helping them set and achieve their goals, unblock them and advocate for their needs within MATS. While RMs can occasionally provide research supervision, this guidance primarily comes from mentors. Although during the extension there can be more scope for this kind of contribution.

I’m also trying to challenge their plans, their theory of change and their conclusions. This comes less naturally to me, requiring a conscious effort to push back, as I remind myself that being overly deferential ultimately fails the scholar. For instance, providing excessive support and insufficient critical feedback early in a project can steer a scholar toward an unfruitful path, such as running uninformative experiments.

Most of this is downstream of a genuine curiosity about the scholars and their research, as this can motivate questions that are also useful for the scholar. However, in my experience, a deep technical understanding of their work isn't always necessary. You’d be surprised how much you can contribute in some cases when just discussing “The Project With Alice” in the abstract, by prompting the scholar to think about prioritisation, collaboration or their own uncertainties. 

Research Management Services

There’s a range of services I aspire to offer scholars as an RM, I’ve tried to list the main ones below. Of course, every scholar has different needs and I don’t claim to provide all of these services. In some cases, I’ve found the RM:Scholar relationship can be fairly minimal weekly progress check ins.

  • Project Management & Accountability
    • Research project planning
    • Weekly goals
    • Daily standups / async updates
  • Life Coaching & People Management
    • Well-being checks
    • Coaching towards longer-term goals
    • Productivity debugging
  • Research Supervision (“pseudo-mentorship”)
    • Giving steer/feedback on research direction
      • Whether the research direction is tractable
      • Whether the current direction serves the ultimate goals of AI safety
    • Providing feedback on experiments to run
  • Research Paper Support
    • Reviewing papers and blog posts
  • Career Coaching
    • Providing guidance on different options available
    • Help with applications, both for jobs and future funding
    • Interview prep / mock interviews
  • Professional Development & Technical Upskilling
  • Networking
    • Connect scholars with collaborators and papers
  • Logistics
    • Helping scholars navigate MATS infrastructure, troubleshooting and generally acting as an interface between MATS and scholars.

Research Manager’s RPG Skilltree

Similarly, RM’s strengths vary. One framing we’ve used internally to reflect on individuals’ strengths is the idea of an RPG-style skill tree for RMs, and how we’d see our 15 skillpoints spent across the major axes of the role.

For the curious here’s my self-assessment:

  1. People Management     ⬤⬤⬤⬤⬤⬤
  2. Project Management    ⬤⬤⬤
  3. Life Coaching                   ⬤⬤
  4. Research Supervision   ⬤⬤⬤⬤

It’s a fairly generalist role and others have given very different allocations. I encourage people to look at the profiles of MATS RMs for a sense of how varied it can be.

I think part of the metaskill of being an RM is having self knowledge of these strengths and being able to quickly identify the jagged frontiers of scholar needs to find overlaps.

What do I actually do?

Currently, I spend less than half my time on direct research management. I’ve always taken on quite a lot of operational / internal project responsibility in London but recently as a senior RM I also now have other RMs to manage, so I now manage half the scholars I used to.

As a result my week is a bit atypical, but I’ll try and outline the major themes of the role below. This is shaped by my experience, but somewhat generalised.

Research Management for Scholars

I consider the RM's core responsibility to be to their scholars.

  • Weekly 1:1 meetings (~30 mins)
    • Check in on progress and well-being
    • Talking through research updates, planning future experiments, red-teaming proposals
    • Setting goals for the next week
    • Preparing for mentor meetings (e.g. talking through uncertainties and what input they need)
    • Navigating MATS
      • Help prepare for milestones (e.g. Scholar Research Plans, Symposium)
      • Support them with MATS’ systems and processes (e.g. compute administration, reimbursement policies)
  • Follow ups (~60mins / scholar, lumpy)
    • Reviewing scholars’ draft outputs and providing comments:
      • Project plans / blog post / paper etc..
    • Refining meeting notes for mentors
    • Reviewing / support with applications for grants / jobs / PhDs
    • Resolving logistical issues
    • Advocating for scholars’ interests
    • (this is where other RMs might provide specialised support such as engineering support, code review etc.)

Research Management for Mentors

Here's a list of things RMs can do for mentors:

  • Summarise meeting notes (removing anything sensitive) / distilled progress update
  • Help scholars preparing demos / communicating results to mentors
  • Planning for milestones within the stream
  • Running / participating in project management sessions

Often the value I am to provide for mentors is to enable their relationship with scholars to focus more on the research, whilst we take care of the other stuff.

Research Management Management

As a senior RM I also manage other RMs, so part of my role is supporting them.

  • People management of RMs
    • Weekly 1:1 check ins
    • Handling escalated questions / uncertainties from RMs (i.e. situations with their scholars, mentors)
    • Generally unblocking & supporting other RMs
  • RM Exchange sessions
    • Running weekly meetings to talk through (anonymised) examples of difficult/interesting situations to collectively review with the London team
  • Supporting hiring, onboarding & training

MATS Projects & Infrastructure

I also contribute to other projects and improvements; there’s a strong ownership culture around projects at MATS.

  • Running sprints for MATS internal projects in London
    • Daily standup meetings
    • Sprint Planning
    • Sprint Retrospectives
  • Events & Milestones in London
    • Planning and running events for scholars in London:
      • Research talks, socials, seminars, workshops, lightning talks, lab visits
    • Milestones for the extension scholars
  • Operational support / projects
    • Setting up a catering system
    • Coordinating with LISA
    • Minor improvements to MATS infrastructure/systems (e.g. creating systems for emailing, collecting references, updates to survey forms etc..)
    • Drafting internal policies
  • Strategic projects
    • Finding new office space
    • Supporting hiring & training in London
    • Contributing to grant applications
    • Writing retrospectives & internal strategy memos

These are some of my projects, but project work varies a lot from RM to RM, based on their strengths.

Reflections on the role

I’ve loved working as an RM, I find it deeply rewarding and I think it is a highly impactful role. Working with MATS scholars was such a privilege, they are just some of the most brilliant and conscientious people earnestly working to solve some of the world’s biggest, wicked problems. I find it extremely rewarding to support them in whatever ways that I can. 

While it remains varied and interesting, I find the internal project work is often less directly rewarding than working with scholars. I felt somewhat conflicted about shifting my focus away from that, but I do think that some of the other (internal project) work I did at MATS is potentially more impactful and that I might have some comparative advantage there, so it seemed important to work on.

Deciding to leave MATS was one of the toughest choices I’ve had to make. However I remain committed to research management within the Alignment space and am excited to continue it at AISI.

7

0
0

Reactions

0
0

More posts like this

Comments1
Sorted by Click to highlight new comments since:

Executive summary: Drawing on personal experience as a London-based Research Manager at MATS in 2025, the author reflects on research management as a generalist, service-oriented role combining scholar support, mentoring enablement, people management, and internal projects, concluding that it is highly rewarding and impactful despite trade-offs that ultimately motivated a transition to AISI.

Key points:

  1. The author frames research management as “servant leadership” plus “radical candor,” focused on helping scholars set goals, unblock progress, and receive timely critical feedback without replacing mentor supervision.
  2. Deep technical expertise is often less important than curiosity, prioritisation support, and challenging scholars’ theories of change early to avoid unproductive research paths.
  3. The RM role spans project management, life coaching, pseudo-mentorship, research output review, career coaching, networking, and logistical navigation, with intensity varying by scholar.
  4. An internal “RPG skill tree” is used to reflect how different RMs allocate strengths across people management, project management, life coaching, and research supervision.
  5. As a senior RM, the author spent under half their time on direct scholar support, with the remainder split across managing other RMs, internal projects, infrastructure, events, and strategy work.
  6. Although internal projects felt less directly rewarding than scholar work, the author judged some to be higher impact given comparative advantage, and remains committed to research management in AI alignment after leaving MATS.

 

 

This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, and contact us if you have feedback.

Curated and popular this week
Relevant opportunities