Mahan Tajrobehkar

Operations Research PhD (UC Berkeley, 2023) and research data scientist; IMO 2013 silver medalist for Iran; former Research Scientist at the credit asset manager Theorem LP.

Notes13

• Quant employment verification: CONFIRMED. The only quant-finance-related employer found is Theorem LP (theoremlp.com), a credit-focused investment manager / asset manager that runs a $2B+ whole-loan and ABS portfolio using ML and data science (backed by Y Combinator and Two Sigma Ventures). Mahan Tajrobehkar held an actual role there: Research Scientist (per LinkedIn, 2023-2025) and Research Intern (2022), and the UC Berkeley IEOR department's official 2023 job-market placement page lists him as 'Researcher at Theorem' - primary-source department confirmation of employment. Classification: this is quantitative finance at an asset manager (credits/loans), NOT a proprietary trading, market-making, or systematic/algorithmic-trading firm; no such role was found. Student programs, competitions and offers are excluded from this conclusion.2023
• PhD in Operations Research, UC Berkeley, 2023; dissertation 'Application-Driven Development of Computational Tools and Algorithms for Machine Learning and Mean-Field Games' (advisor-side work with Xin Guo).2023
• Silver medal for Iran at the International Mathematical Olympiad 2013 (score 61, 88.59%, 28 points) - the Olympiad identity anchor that ties this research to the correct person.2013
• Current role: Research Data Scientist at Google (from 2025), working on 'GenAI for safety and safety for GenAI' per LinkedIn - Google is a technology company, not a quantitative-finance employer.2025
• Co-authored 'Utilization Analysis and Fraud Detection in Medicare via Machine Learning' (medRxiv preprint, 2024) with Xin Guo et al. - applied ML work dated after the PhD.Dec 2024
• MFGLib: A Library for Mean-Field Games (arXiv 2304.08630, 2023) - co-authored open-source software library; he is listed with a berkeley.edu email ([contact omitted]).2023
• Theorem work was technical R&D, not trading: the LinkedIn role description is 'ML-Driven credit underwriting', consistent with Theorem's ML/data-science credit-risk product rather than portfolio trading.2023
• Published the MFGLib software paper in JOSS (Journal of Open Source Software), the peer-reviewed venue for research software.2023
• Research Intern at Theorem, San Mateo, California, 2022 (3-month summer internship) - an actual internship, distinct from the later full-time Research Scientist role.2022
• 'Escaping Saddle Points Efficiently with Occupation-Time-Adapted Perturbations' (Guo, Han, Tajrobehkar, Tang) - optimization research from his Berkeley period.May 2020
• Education: Bachelor's degree in Electrical and Electronics Engineering, Sharif University of Technology, 2013-2018, before the UC Berkeley graduate work.2018
• Google Scholar profile (verified @berkeley.edu email) lists research areas convex/nonconvex optimization and machine learning.2026