David Matolcsi

Dávid Matolcsi is a Hungarian mathematician and AI safety researcher. He represented Hungary at the IMO in 2018 (bronze) and 2019 (silver), then studied mathematics at Eötvös Loránd University (ELTE). He was a Research Intern at WorldQuant in Budapest in 2022, a Researcher at the Alignment Research Center (ARC) in Berkeley from 2023 to 2025, and since 2025 has been a Technology Specialist at the European AI Office in Brussels.

Notes10

• Quant employment verification: UNRESOLVED. The only quantitative-finance role on record is a Research Intern at WorldQuant, Budapest, Jun 2022-Sep 2022, which is an actual internship (not a student program) at a firm whose business is quantitative/systematic investment (WorldQuant LLC, an affiliate of Millennium). However, this role is sourced only from the LinkedIn index row, not from an independent primary source (WorldQuant career page, press, or a co-worker), so actual employment cannot be independently corroborated; identity as the IMO 2018/2019 Hungarian contestant is supported by that same index profile linking the WorldQuant role together with the ELTE mathematics education. No other quant/trading/market-making/hedge-fund/bank/asset-management/energy/crypto trading role or internship was found; no student insight days, camps, or scholarships at quant firms were found either. Student competitions (IMO, IMC) and the MATS program are not employment and do not qualify as quant work.Jun 2022
• Represented Hungary at the International Mathematical Olympiad twice, earning bronze in 2018 and silver in 2019.
• Co-author of 'Ctrl-Z: Controlling AI Agents via Resampling' (arXiv 2504.10374, 2025), an AI-control paper credited with writing contributions.Apr 2025
• Technology Specialist in the AI Safety Unit of the European AI Office (European Commission), Brussels, since 2025; the MATS alumni page independently lists him as 'AI Safety Unit at the AI Office of EU commission'.2025
• Studied mathematics at Eotvos Lorand University (ELTE) in Budapest; the LinkedIn index lists ELTE as his education, consistent with his ELTE IMC team membership in 2021.
• Researcher at the Alignment Research Center (ARC), Berkeley, 2023-2025, per the LinkedIn index; ARC is an AI-safety research nonprofit, not a quantitative-finance firm.
• Listed co-author of 'Chain of Thought Monitorability: A New and Fragile Opportunity' (arXiv 2507.11473, 2025).2025
• Worked on tiling and weak tiling in finite abelian groups (Fuglede/spectral-set problems), joint work with G. Kiss and Mate Matolcsi of the Renyi Institute, presented at Harmonic and Spectral Analysis 2023.2023
• Scored 56 points at the 2021 International Mathematics Competition for University Students (IMC) as part of the ELTE team, shared 11th-15th place.Aug 2021
• Google Scholar profile carries a verified email at the Renyi Institute (matolcsi domain), consistent with combinatorics research ties there; no public email address string could be cited.
• Listed alumnus of the MATS (Machine Alignment/Technical Safety) program, an AI-safety research training program; this is a scholarship/program, not employment.
• Maintains a public LessWrong account (david-matolcsi) with a shortform post; no X/Twitter or GitHub account is linked from it.