PhD Positions in AI Safety

PhD Positions in AI Safety – University of Stuttgart

Join the independent AI Safety research group headed by Dr. Thilo Hagendorff and apply for four PhD positions (TV‑L 13, 100 %). The research focuses on authenticity, deception, and situational awareness in large reasoning models.

Responsibilities

• Conduct high‑impact empirical research on large reasoning models and safety.
• Design, evaluate, and publish safety evaluation methods and data sets.
• Collaborate with computational‑linguistics, interpretability, and monitoring experts.
• Contribute to dissertation, journal papers, and conference presentations.

Qualifications

• A Master’s degree (or expectation of completion) in computer science or a related field.
• Familiarity with large‑scale reasoning models and large‑language models.
• Strong analytical and programming skills (Python, statistics).
• Empirical research mindset and willingness to learn new skills (interpretability, monitoring).
• Advanced English language skills (German not required).
• Interest in empirical evaluation of AI systems, especially deception and situational awareness.

Offer

PhD candidates receive support from ELLIS Stuttgart, GRADUS, and IRIS. The contract is TV‑L 13. Two projects offer two contracts of up to 2 years and two contracts of up to 1.5 years. The envisaged start date is 01 01 2026 but can be earlier.

How to Apply

Please submit a single PDF application to by 15 November 2025, including: cover letter, CV, publication list, a document with your most important achievements, favourite papers and why, main limitations of current AI models, certificates and reference letters (if available).

Equal Opportunity Statement

The University of Stuttgart encourages applications from women, severely disabled persons, and other under-represented groups. German or English diplomas may be required.

#J-18808-Ljbffr