Two fully funded three-year research positions, one PhD and one postdoc, have opened at the CopeNLU group at the University of Copenhagen. The project tackles how LLMs absorb, propagate, and resist false information at the intersection of mechanistic interpretability, LLM security, and explainable AI. Both positions start in September 2026. The deadline is 31 May 2026.
24 days to the deadline. If you need transcripts, references, or visa guidance, start today.
About the Programme
CopeNLU is the Natural Language Understanding research group at the University of Copenhagen’s Department of Computer Science. Professor Isabelle Augenstein and Assistant Professor Pepa Atanasova lead it.
The two positions are part of a new project: A Mechanistic Framework for Mitigating the Susceptibility of LLMs to Learning False Information, funded by the Independent Research Foundation Denmark. The project will develop theoretical frameworks for LLM security, new mechanistic interpretability methods, and new evaluation protocols, all at the intersection of NLP, LLM security, and explainable AI.
The project includes a separate pathway to apply as an academic collaborator with NVIDIA. This is not automatic; it requires its own application as part of an existing institutional relationship. Scope and access terms are not detailed in the listing.
See Also: ITCILO Internship 2026: Data and AI Role Reserved for Candidates With Disabilities
What You Will Research
PhD position: The PhD work centres on how LLMs become vulnerable to false information at each stage of their lifecycle, from pre-training through deployment, and what mechanistic interventions can reduce that vulnerability. The PhD student will be supervised by Augenstein and co-supervised by Atanasova.
Postdoc position: The postdoc maps how false information enters and moves through LLMs at each lifecycle stage. From that mapping, the postdoc builds the defences: mitigation methods and attack-prevention protocols. The postdoc will work directly with both PIs and the wider project team.
Both positions sit at the intersection of NLP, LLM security, and explainable AI. These three fields are rarely tackled together. Their convergence matters as language models move into high-stakes settings.
Salary and Benefits
PhD salary: The University of Copenhagen’s standard PhD pay scales set the monthly base at approximately DKK 25,000 before tax. The university contributes 17.1% toward pension, bringing the total monthly package to approximately DKK 30,800. LLM security and mechanistic interpretability matter; both remain underfunded relative to how widely language models are now deployed, making fully salaried positions in this area uncommon.
PhD students also receive an annual research budget for conferences, books, and travel, plus a computer, office phone, and office space.
Postdoc salary: Not specified in the listing.
Both positions run for three years and start in September 2026, or as soon as possible thereafter.
Eligibility
PhD position
- A completed Master’s in Computer Science or a relevant discipline
- Strong background or prior experience in ML or NLP
- Candidates with a master’s still in progress may apply if completion is expected before September 2026; confirm this directly.
Postdoc position
- A completed PhD in Computer Science or a related area
- Research or work experience in NLP, LLM Security, and/or Explainable AI
- Candidates who have submitted but not yet received their PhD may still be eligible; confirm with the PIs directly.
Both positions
- Strong English communication skills
- Whether TOEFL or IELTS scores are required is not stated in the listing.
- Non-EU applicants will need a Danish work and residence permit before employment begins; the process typically takes six to twelve weeks.
Documents Required
- CV or academic resume
- Cover letter connecting your work specifically to this project’s goals
- Degree certificate and transcripts (PhD applicants) or PhD certificate (postdoc applicants)
- Contact details for two or three academic referees
- A writing sample: thesis chapter, published paper, or preprint
Reference letters and official transcripts take the longest to gather. With 24 days to the deadline, contact your referees and transcript office now.
Application Timeline
You have 24 days. Use them in order.
- Now: Email your referees. Request official transcripts. Read the PIs’ recently published work.
- By 14 May: Draft your cover letter. Map your experience directly to the project goals, such as mechanistic interpretability, LLM security, and evaluation protocols.
- By 21 May: Select your writing sample. Finalise your CV and research statement.
- By 28 May: Confirm references are submitted. Review all documents. Submit early; do not wait for 31 May.
- Non-EU applicants: Begin your visa inquiry now. September 2026 is 16 weeks away. Processing takes 6–12 weeks.
See Also: MTN Foundation Scholarship 2026: ₦300,000 a Year Until You Graduate
Is This Open to Nigerians?
CopeNLU maintains an international and inclusive group. The listing states no regional restriction. Nigerian and African applicants are not excluded.
Non-EU nationals will need a Danish work and residence permit. Universities typically support this process, but the timeline matters for a September 2026 start. If you are applying from outside the EU, initiate the visa enquiry before you submit your application, not after.
How to Apply
PhD position
- Read the full listing
- Prepare your CV, cover letter, transcripts, writing sample, and references.
- Connect your prior work specifically to mechanistic interpretability or LLM robustness in your cover letter; a generic NLP interest will not distinguish you.
- Submit via the University of Copenhagen employment portal by 31 May 2026
Postdoc position
- Read the full listing
- Prepare your CV, publication list, research statement, and references.
- Your research statement should address the project’s three goals directly: false information characterisation, mitigation methods, and evaluation protocol development.
- Submit via the University of Copenhagen employment portal by 31 May 2026
GizPulse Verdict
Among funded NLP research positions available globally right now, this one sits near the top. LLM security and mechanistic interpretability matter. Both remain underfunded relative to how widely language models are now deployed. A three-year fully funded position at a top European lab, co-supervised by two active researchers in exactly this area, with a NVIDIA collaboration pathway available. Positions like this do not appear often.
Copenhagen’s cost of living is high, but PhD salaries are structured to match it.
Who this is for: An NLP researcher with strong ML foundations who has done or wants to do work on model internals, robustness, or information integrity. The strongest PhD applicants will have a master’s thesis touching on at least one of the following: interpretability, adversarial attacks, fact-checking, or misinformation. The strongest postdoc applicants will have publications in NLP security or explainable AI.
Who should look elsewhere? Researchers focused on generative AI applications or product-facing NLP. This is blue-sky academic research; the outputs are papers and frameworks, not deployed products.
One tip: Both PIs publish actively in exactly the areas this project addresses. Read two or three of their recent papers before writing your cover letter and reference specific methodological choices or open questions they have raised. That separates genuine candidates from bulk applicants.
Stay ahead of the best research fellowships, PhD openings, and opportunities across Africa’s tech sector. The GizPulse Weekly delivers them to your inbox.
