Mark Jabbour
Mark Jabbour (Syria) competed at the IMO in 2016, 2017 and 2018, earning Bronze, Silver and Bronze medals. He studied Mathematics and Computer Science at MIT (BS 2019-2023, with a further MIT year through 2025) and works as an ML Research Engineer at Windsurf. His professional experience includes quantitative research internships at the hedge fund Point72 and the systematic trading firm Vatic Labs, and a software engineering internship at the proprietary trading firm Jump Trading. He co-authored a 2026 arXiv paper on evaluating frontier language models on physics problems.
- Quant employment verification: CONFIRMED. Supporting evidence: index (LinkedIn) records actual internships, not student programs - Quantitative Research Intern at Point72 (New York, 2024), Quantitative Research Intern at Vatic Labs (New York, 2023), and Software Engineering Internship at Jump Trading Group (Chicago, 2022). Employer classification: Point72 is a multi-strategy hedge fund with a systematic-quant arm (Cubist Systematic Strategies); Vatic Labs is a quantitative/systematic trading firm; Jump Trading is a proprietary high-frequency trading/market-making firm - all genuinely quantitative-finance employers. Identity is supported: the same profile carries the Syrian IMO background and MIT Mathematics/CS education, corroborated by independent MIT-institution records (arXiv co-authorship 2026, MIT 2025 commencement). Caveat retained: the employment entries themselves are index-only (single-source LinkedIn index); no independent primary employer page or press announcement confirming these internships was located, so the roles are asserted as listed rather than independently confirmed.
- Identity: the Syrian IMO contestant Mark Jabbour matches the MIT Mathematics/Computer Science graduate and Windsurf ML Research Engineer on LinkedIn profile mark-jabbour-4a743b1a6 (name + Syrian IMO + MIT alignment). Identity is supported by concordant index data plus independent MIT-institution records (arXiv paper co-authorship, MIT 2025 commencement listing), but the LinkedIn profile itself is an index source, not a primary identity document.
- Index-only LinkedIn data lists three quantitative-finance internships - Quantitative Research Intern at Point72 (New York, 2024), Quantitative Research Intern at Vatic Labs (New York, 2023), Software Engineering Internship at Jump Trading Group (Chicago, 2022). These are actual internships (not student programs/insight days) but rest on a single index source; no independent primary employer page confirming them was located.
- Co-authored the arXiv paper 'How Good Are Frontier Models at Physics? Expert Re-Grading Reveals Broken Evaluations and Near-Saturation of Leading Benchmarks' (arXiv:2609.13009, cs.AI, 11 Sep 2026) alongside Yale/Cambridge/USC physicists - shows active ML-evaluation research output, consistent with the Windsurf ML Research Engineer role.Sep 11, 2026
- ML Research Engineer at Windsurf (2025-): per the role description, contributed heuristics and LLM-based methods to find interesting subsets of user data and curate evaluation/training datasets for the SWE-1 model.Jan 1, 2025
- Represented Syria at the IMO three times per the official imo-official.org record: 2016 (rank 206, 66.0%, 18 pts, Bronze), 2017 (rank 82, 86.8%, 21 pts, Silver), 2018 (rank 228, 61.8%, 18 pts, Bronze) - an academic competition record, not employment.Jul 1, 2018
- Employer classification: Point72 is a large multi-strategy hedge fund (owner of Cubist Systematic Strategies), Vatic Labs is a quantitative/systematic trading firm, and Jump Trading is a proprietary high-frequency trading and market-making firm - all genuinely operate in quantitative finance, so the names are not merely 'quant'-sounding.
- MIT record: listed in the MIT Class of 2025 commencement name book with thesis title 'Transformers as Empirical Bayes Solvers', and the LinkedIn education block shows MIT 2019-2023 (BS, Mathematics and Computer Science) plus a further MIT year 2024-2025 - suggests an advanced/MEng year at MIT awarded 2025.May 30, 2025
- Listed as 'Mark Jabbour - @Windsurf' among guests on a GTC 2025 AI Agents Hackathon event page, tying his public footprint to the Windsurf role in 2025.Mar 1, 2025
- Taught an ESP MIT Spark 2020 course on Sperner's Lemma (instructor alongside Arvid Lunnemark), an academic-outreach/teaching data point preceding his professional roles.Mar 1, 2020
* Contributed to a set of heuristics and LLM-based methods to find interesting subsets of user data * Contributed to curating evaluation and training datasets for SWE-1 * Contribut
Competition record
Syria · IMO