Project Lead, Technical AI Governancefrom January 2027
AI Safety Camp 12
Selected to lead a team modeling compute-permit prices and audit rules.
Abstract
I'm a technical AI safety and governance researcher. My work sits between the technical side of AI and its regulation and governance. I have a PhD in computer science from the University of Tübingen, on privacy-preserving machine learning, and I hold the AIGP and CIPP/E certifications in AI governance and data protection. Before that, I studied mathematics and mathematical physics.
Keywords: compute governance; third-party auditing; agentic AI protocols; black-box attacks.
1Selected publications
- [1]PP-GWAS: Privacy Preserving Multi-Site Genome-wide Association StudiesNature Communications, 2025. Code
- [2]Position: Trustworthy Model Context Protocol Enables Responsible Agentic AI!ICML, 2026. Position paper.
- [4]Risk-Weighted Compute Permits: Enforcement Design Under Imperfect MonitoringUnder review.
- [5]Empirical Evidence for Simply Connected Decision Regions in Image ClassifiersarXiv, under review. Code
2Current work
Independent Researcher, Technical AI Safetysince October 2026
Full-time research and training in technical AI safety.
Research Fellow, Technical AI Governancesince September 2026
Supervised Program for Alignment Research (SPAR)
Studying how agent interoperability standards affect intentional loss-of-control risk.
MentorKevin Kohler
3Past work
Research Fellow, Technical AI GovernanceApril to September 2026
Safe AI Germany (SAIGE)
Extended a formal model of risk-weighted compute permits to audits facing hidden compute and manipulated evaluations. Led the resulting paper and co-authored a public policy brief.
MentorJoel Christoph
Scientific Researcher, Privacy-Preserving Machine LearningOctober 2022 to September 2026
University of Tübingen
Eight publications. Built and open-sourced end-to-end implementations of privacy-preserving methods.
Teaching, tutoring and supervision2022 to 2026
Taught and mentored 75+ graduate students. Supervised a master's thesis through to publication.
PhD RepresentativeJune 2023 to June 2026
Represented 60+ PhD students from nineteen research groups at the Institute for Bioinformatics and Medical Informatics, University of Tübingen.
4Education
| 2022 to 2026 | PhD, Computer ScienceUniversity of TübingenThesis: Privacy by Design: From Distributed Learning to Post-Deployment RisksGrade: 1.0 (100%), magna cum laudeSupervisors: Dr. Mete Akgün and Prof. Dr. Nico Pfeifer |
|---|---|
| 2020 to 2022 | MSc, Mathematical PhysicsUniversity of TübingenThesis: Supersingular Isogeny Diffie-Hellman Key ExchangeGrade: 1.53 (91%)Thesis supervisor: Prof. Dr. Hannah Markwig |
| 2016 to 2019 | BSc, MathematicsAmrita UniversityThesis: Resolution of Plane Curve Singularities: Blowing UpGrade: 1.4 (92%), First Class with Distinction |
5Certifications
6Grants and awards
- Career Transition Grant, Coefficient Giving, since October 2026.
- Leadership Talent Academy Scholarship, Startup Center, University of Tübingen, 2025. One of twenty recipients of a three-month leadership and venture-building program.