Leaders of TodayLeaders of Today
BrowseWorkspaceProfile
Leaders of Today

Every opportunity worth finding, checked against its source.

Submit an Opportunity
Leaders of TodayAbout UsBlogContactReport a Bug
LegalCookiesDisclaimerPrivacy PolicyTerms of Service

© 2026 Leaders of Today™. All Rights Reserved.

This opportunity has closed. It stays published for reference, and new cycles reopen here when the organizer announces them. Browse open opportunities

€5,000 AI Safety Research Fellowship

  • Organization: Odyssean Institute
  • Type: Fellowship
  • Location: Global
  • Deadline: September 11, 2026
  • Funding: Fully Funded

The Odyssean Institute, a UK research charity working on AI and other catastrophic risks, is recruiting three researchers for the first cohort of its research fellowship. Over three months you will help build a method for turning what AI safety experts believe into causal models other people can inspect and argue with, using LLMs to code expert testimony and systems modelling to map it. The work is remote, paid, and published openly as an open-source project. **What you get:** - A stipend of 5,000 euros for the three-month engagement - A place in the inaugural cohort of three fellows, shaping a new programme from the ground up - Remote working, with the option of hybrid collaboration if you are in the UK or the USA - Named contributions of code, documentation and methodology write-ups to an open-source repository - Research discussions with a transdisciplinary team whose backgrounds include Oxford, Cambridge, the Australian National University, the National University of Singapore and Goldsmiths **What you'll do:** - Develop and refine LLM-assisted pipelines for thematic coding of qualitative expert data - Extract the causal variables and hypothesised relationships experts see in how AI safety unfolds - Build, iterate on and document causal loop diagrams and, where appropriate, stock-and-flow models - Contribute code, documentation and methodology write-ups to an open-source repository - Help design and support expert elicitation sessions and workshops where relevant - Submit a weekly timesheet tracking hours and progress **Key dates:** - Applications close 11 September 2026 - Fellowship starts 14 September 2026 - Three-month engagement at 20 hours per week or more **More information:** - Apply by sending a single PDF to operations@odysseaninstitute.org and contact@odysseaninstitute.org. There is no application form - The application asks for three things: a one-page sketch of a first small experiment on one of the four fellowship tasks, evidence of previous thematic or qualitative coding work with a note on your specific role, and a few paragraphs on how you would explore a causal loop diagram under deep uncertainty - The Institute asks applicants to limit AI use to review and grammatical edits, and says honest partial answers are welcome because it is reading for judgment rather than polish - The Odyssean Institute is a registered UK Charitable Incorporated Organisation, charity number 1204794

Apply for €5,000 AI Safety Research Fellowship

Odyssean InstituteFellowship

€5,000 AI Safety Research Fellowship

This one has closed
Archived
Opportunity type
Fellowship
Funding
Fully Funded
Location
Global
Deadline
Archived

About this opportunity

The Odyssean Institute, a UK research charity working on AI and other catastrophic risks, is recruiting three researchers for the first cohort of its research fellowship. Over three months you will help build a method for turning what AI safety experts believe into causal models other people can inspect and argue with, using LLMs to code expert testimony and systems modelling to map it. The work is remote, paid, and published openly as an open-source project.

What you get:

  • A stipend of 5,000 euros for the three-month engagement
  • A place in the inaugural cohort of three fellows, shaping a new programme from the ground up
  • Remote working, with the option of hybrid collaboration if you are in the UK or the USA
  • Named contributions of code, documentation and methodology write-ups to an open-source repository
  • Research discussions with a transdisciplinary team whose backgrounds include Oxford, Cambridge, the Australian National University, the National University of Singapore and Goldsmiths

What you'll do:

  • Develop and refine LLM-assisted pipelines for thematic coding of qualitative expert data
  • Extract the causal variables and hypothesised relationships experts see in how AI safety unfolds
  • Build, iterate on and document causal loop diagrams and, where appropriate, stock-and-flow models
  • Contribute code, documentation and methodology write-ups to an open-source repository
  • Help design and support expert elicitation sessions and workshops where relevant
  • Submit a weekly timesheet tracking hours and progress

Key dates:

  • Applications close 11 September 2026
  • Fellowship starts 14 September 2026
  • Three-month engagement at 20 hours per week or more

More information:

  • Apply by sending a single PDF to operations@odysseaninstitute.org and contact@odysseaninstitute.org. There is no application form
  • The application asks for three things: a one-page sketch of a first small experiment on one of the four fellowship tasks, evidence of previous thematic or qualitative coding work with a note on your specific role, and a few paragraphs on how you would explore a causal loop diagram under deep uncertainty
  • The Institute asks applicants to limit AI use to review and grammatical edits, and says honest partial answers are welcome because it is reading for judgment rather than polish
  • The Odyssean Institute is a registered UK Charitable Incorporated Organisation, charity number 1204794

Who can apply

  • Mixed-methods research experience, with hands-on familiarity with thematic coding of qualitative data
  • Scenario development and analysis
  • Complex systems analysis methods, particularly causal loop diagrams and stock-and-flow modelling
  • Experience using LLMs to build open-source projects, including prompting, orchestration and evaluation of outputs in a working pipeline
  • Comfort with methodological experimentation and ambiguity, and willingness to discard what does not work and iterate
  • Available for 20 hours per week or more across the three months from 14 September 2026
  • Applicants are asked to attempt all three application questions, and to limit AI use to review and grammatical edits
  • No age limit, nationality requirement or degree requirement is stated in the listing
  • Researchers with hands-on mixed-methods experience, including thematic coding of qualitative data, plus working knowledge of complex systems methods such as causal loop diagrams and stock-and-flow modelling. The Institute also asks for experience using LLMs to build open-source projects, covering prompting, orchestration and evaluation of outputs in a working pipeline, and for scenario development and analysis. Three fellows are selected so their combined skills cover that range, and the listing states you do not need every one of these. Quantitative risk estimation, expert elicitation methods such as the IDEA Protocol, prior AI safety or global catastrophic risk research, and past open-source contributions are listed as nice to have. This is a genuinely experienced-researcher brief rather than an entry-level one, but no age limit, nationality requirement or degree requirement appears anywhere in the listing.

AI Verification

✓ AI Verified

Our AI read the official sources and cross-checked the key facts. Here is what held up, and what it could not confirm.

We could not fully verify this organization or all of its details. Check it independently on the official website before you apply, and never share sensitive personal or financial information until you are sure a request is genuine.

69
LOT Verification ScoreFair

Cross-checked 18 of 21 key facts

Confirmed Our best read Not stated

Good to know

  • +EUR 5,000 for three months at 20 hours a week, so it fits around other work
  • +Free to apply. It is a single PDF emailed to the Institute
  • +Remote, with no residency requirement and no travel cost on the fellow
  • +An inaugural cohort of three, run by a UK registered charity

Worth checking yourself

  • !Applications close 11 September and the fellowship starts on 14 September, three days later. This is now or not at all
  • !The only information page is a Google Doc, because the Institute has never published one on its own site
  • !Hybrid working is offered only to people already in the UK or the USA

What we could not confirm

  • ?Deadline Time: the listing states 11th September with no clock time. The 23:59 house default has been applied
  • ?Deadline Timezone: no timezone is named anywhere in the listing. UTC recorded as the documented default; the organisation is UK based, so the true cutoff is most likely 23:59 UK time
  • ?Deadline year: the listing writes 11th September with no year. 2026 is confirmed by the 3 September 2026 announcement post, the 14 September start date, and the live Probably Good listing dated Sep 11, 2026

Sources checked

ODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialODodysseaninstitute.orgOfficialDOdocs.google.comApplyJOjobs.probablygood.orgOfficialFOforum.effectivealtruism.orgOfficialFOforum.effectivealtruism.orgOfficialMAmanifund.orgOfficialCOcordademocracy.orgOfficialPRprobed and 404: /fellowship /fellowships /research-fellowship /careers /jobs /join-us /opportunities /get-involved /vacancies /work-with-us /apply /fellows on odysseaninstitute.orgOfficialWAWayback CDX index for odysseaninstitute.org, which shows no fellowship, apply or careers page has ever been archived on the domainOfficial

LOT finds and checks opportunities from public sources. Always confirm details on the organiser’s own page before applying.

€5,000 AI Safety Research Fellowship
Archived
This one has closed