SERIES · This post collects practical advice and changes for the upcoming applications to postdocs and early career jobs. If you come across other advice published by hiring committees this year, or if you are involved in hiring and would like to add something, please tell us: we will add it to this post. All posts in the series on careers
Cambridge: three changes to postdoc appointments at DPMMS
DPMMS has adapted the appointments process for the Herchel-Smith (and other departmental) postdoc in three ways this season. First, we ask candidates in their applications to specify up to four of their strongest papers or preprints, to consciously de-weight volume of production relative to quality. Second, we have a small bespoke hiring committee to explicitly compare files across different fields, which includes taking stock of different experiences around AI which is impacting fields differently. Finally, we have told applicants we expect to interview for the positions, which we have not done before, giving leading candidates a chance to show their command of and passion for their work more directly.
Received 11 November 2026.
CNRS: advice from the Section 01 committee for the 2027 junior researcher competition
In 2027, the CNRS should open 13 junior researcher positions in mathematics. These are tenured research-only positions, 9 of them based at any French university and 4 of them reserved for particular universities; in the past few years, the average academic age of new hires in mathematics has been PhD + 3/4 years, and we welcome international applications. (No knowledge of French is necessary to apply or be selected.)
The CNRS mathematics hiring committee has published some advice to candidates to explain our position concerning the use of generative AI during the hiring process, the changes we are making to adapt to the new circumstances, and our evaluation criteria. This advice and more information about the calendar is available on our website: French version here, English version here. Here is a short summary:
- We will not use any generative AI for our work as a committee. This means in particular that we will not put any part of any application in any generative AI tool, and we will not use generative AI tools to judge applications.
- We will not penalize candidates for using or not using generative AI tools in their research.
- We ask candidates to include a detailed declaration of AI use in their research, and to write their report on previous research and their research project themselves (without assistance from generative AI). We also ask them to make these texts as clear and accessible as possible; they should be understandable both to specialists in the field and to mathematicians working outside the candidates’ area of research.
- The application should not include a full list of publications and preprints. Instead, we ask candidates to select at most 7 scientific productions (articles, preprints, software, databases etc) and to provide a short explanation of their selection.
- In the report on previous research, we ask candidates to highlight at most 3 research contributions and to explain them in detail (context, new ideas and methods, etc).
- We will double the length of the interviews and the size of the interview subcommittees.
Received 13 November 2026.
ETH Zurich: postdoc hiring in the age of AI
Drawing on our experience as past and current members of postdoctoral hiring committees, we would like to share our perspective on how to evaluate postdoctoral candidates in the age of AI.
AI does not change what we ultimately want to assess in a candidate: the depth of their mathematical understanding, the substance of their own contributions, and their potential to develop new ideas.
At the initial shortlisting stage, we expect recommendation letters and CVs to play an important role, but we do not regard publication counts alone as a reliable guide to a candidate’s potential. We are more interested in the substance of the work and in what the application reveals about the candidate’s own contributions, ideas, and development as a mathematician. Where possible, we may also examine selected papers more closely to gain a better understanding of the mathematical ideas involved and how they are presented.
For shortlisted candidates, the interview will be crucial for assessing their mathematical understanding and judgment more directly.
Can they explain the ideas behind their results, why the arguments work, and where the difficulties lie?
Can they discuss what they tried, what failed, and which questions remain open?
We would also like to understand how they see their work in a broader mathematical context, why they consider it significant, and what they regard as their own main contributions. We expect that a talk followed by an in-depth discussion will give us a good sense of each candidate’s mathematical understanding, maturity, and approach to research.
Our advice to applicants is therefore to make their own contributions clear in their applications and to be prepared to discuss their work in depth, since a thoughtful and honest mathematical conversation can reveal far more than a publication count. While the ways in which we assess candidates may evolve as the tools available to mathematicians change, our emphasis will remain on mathematical understanding, originality, and research potential.
Received 5 November 2026. First published on 6 November 2026 as “Postdoc Hiring in the Age of AI: What Still Matters”.
Elsewhere: advice published by other institutions
These notices were not sent to Proofs and Prompts. We quote them as the institutions published them, with a link to the source and the date we consulted it.
Institute for Advanced Study, School of Mathematics
Postdoctoral and other memberships; deadline 1 December. The School’s membership application page includes the following notice (consulted 13 November 2026):
PLEASE NOTE: We realize that applying to positions in the current environment is stressful. To help alleviate this, the School of Mathematics offers applicants the option to submit research statement based only on already conducted research. It would be helpful if such a statement would be focused on a few select projects (in most cases up to three) and not a wide ranging list.
We also want to turn the attention of applicants to the option of including a url to a recorded lecture (e.g. seminar or conference talk) in the application form. If you did not have a talk recorded, you can record one yourself and give the link; if this is not clear from the accompanying information in the video url, please explain in which context the talk was given and to what audience in the comments section of the application.
Applicants should feel free to include in the comments section any pertinent information regarding the application, in particular regarding impact (either helpful or unhelpful) of AI related tools on your research.
IAS School of Mathematics, “Applying for Membership in the School of Mathematics”
Max Planck Institute for Mathematics, Bonn
Guest Program and postdoc positions; the fall 2026 round closed on 16 October. The institute’s application page includes the following AI policy (consulted 13 November 2026):
The MPI is a place for human interactions and discussions.
Your research statement must be fully written without the use of AI, and we encourage you to keep it to at most 5 pages. When describing your previous work, please disclose the extent of AI use therein.
We will not use AI in the evaluation of applications.
The Max Planck Society provides free access to current LLM’s for everybody who is at MPIM for more than a year, and researchers who use AI in their work are welcome.
We believe that it is important to sustain a subcommunity of mathematicians who choose not to use AI in their work, and we also welcome their applications.
We expect our members to support the Leiden Declaration on Artificial Intelligence and Mathematics.
Max Planck Institute for Mathematics, “Application for the Guest Program and for postdoc positions”
MIT, Department of Mathematics
C.L.E. Moore Instructorships (deadline 1 December) and tenure-track faculty positions (deadline 1 November). Both postings on the department’s employment page ask for a “Gen AI statement”, described as follows (consulted 13 November 2026):
The Gen AI statement is a detailed factual statement on the use of generative AI, if any, in the applicant’s past work and current application materials.
MIT Department of Mathematics, “Employment”
Recommendations to hiring committees
These are proposals addressed to departments, not rules of any hiring process.
- Report of the Summit on Math Research in the Age of AI (Harvard CMSA, 22–23 October 2026). Among other things, it suggests that during this transition departments may allow research statements without future research plans, and that application materials could include evidence of exposition and communication.
- AI and TCS: The Next Six Months (Simons Institute working group, 28 September 2026), for theoretical computer science. It recommends evaluating candidates on a few works of their own choosing, testing their understanding of those works in more depth, and making these changes starting with the 2026–2027 hiring season.
Add to the discussion