DFKI Postdoctoral Position in Multilingual Mechanistic Interpretability
German Research Center for Artificial Intelligence (DFKI)
DFKI is hiring a full-time postdoctoral researcher for the three-year SLIMS project on reasoning circuits and mechanistic interpretability across compact multilingual language models and more than 20 languages. Applications close September 15, 2026.
The German Research Center for Artificial Intelligence (DFKI) is recruiting a Senior Researcher/PostDoc in Multilingual Mechanistic Interpretability in Saarbruecken. The full-time position is funded for 36 months, with a planned start of January 1, 2027 and possible extension.
The researcher will join the BMFTR-funded SLIMS project, a bilateral collaboration between DFKI and Inria. The work will systematically study reasoning circuits in compact multilingual language models, including transfer across model scales from roughly 1B to 7B parameters and across 20 to 30 languages. Methods include activation patching, sparse autoencoders, nonlinear attribution and transcoders.
The position sits within DFKI MLT under Marius Mosbach and is primarily supervised by Dr. Simon Ostermann. It includes direct collaboration with Inria Paris and the ALMAnaCH team, including planned research visits and joint workshops.
Employment: DFKI offers competitive, market-based compensation for a three-year full-time contract. The precise salary is not published.
How to apply: submit the official DFKI online application by September 15, 2026, quoting reference 11388-2026. Questions may be sent to Simon.Ostermann@dfki.de.
Source and verification
Check the current official call before applying
This independent listing was last checked on September 1, 2026. We compare the named provider, deadline status, eligibility, funding claims and application destination where those details are available. The live provider page is the final authority.
Benefits
- Competitive, market-based salary for 36 months, with possible extension; exact salary not published.
- Access to DFKI GPU infrastructure and research resources.
- Research visits and joint workshops with Inria ALMAnaCH in Paris.
- Support for publications, conferences and international collaboration.
- Opportunities to mentor and co-supervise students.
- Employee benefits may include Urban Sports Club access, corporate benefits and a Jobticket subsidy.
Eligibility summary
- Open to qualified applicants of any nationality; no nationality restriction is stated.
- Applicants must have completed a relevant PhD and be available for full-time employment in Saarbruecken, Germany.
- DFKI encourages applications from women in science and gives preference to applicants with disabilities when equally qualified.
Requirements
- PhD in computer science, natural language processing, linguistics, machine learning or a closely related field.
- Strong expertise in mechanistic interpretability.
- Research experience in multilingual or low-resource NLP.
- Excellent Python and deep-learning implementation skills.
- Strong publication record at leading research venues.
- Ability to collaborate with DFKI and Inria partners, undertake research visits and help supervise students.
Documents required
- Completed DFKI online application form.
- Curriculum vitae.
- Cover letter or statement of research fit.
- Doctoral degree certificate and academic records.
- Publication list and selected research outputs.
- Any additional documents requested by the DFKI application portal.
Language/test requirements
How to submit applications
Provider: German Research Center for Artificial Intelligence (DFKI)
Official DFKI vacancy reference 11388-2026. Full-time, temporary, 36 months. Planned start January 1, 2027. Deadline September 15, 2026.
Listing-board disclaimer
Prime Scholarship Alerts is an independent listing board, not the scholarship provider, university or funder. We do not select applicants, collect provider application fees or guarantee an award. Verify requirements, deadlines, funding, fees and application instructions through the current provider page before applying.