Alexander Wei

Research scientist at OpenAI (Member of Technical Staff since January 2024) working on large language models and reasoning, with a focus on mathematical reasoning and safety. Ph.D. in Computer Science from UC Berkeley (2023), previously Harvard (A.B. summa cum laude, S.M.). Gold medalist for Team USA at IOI 2015; Summer 2019 quantitative research intern at D. E. Shaw & Co.

Notes5

• Quant employment verification: CONFIRMED. His CV at alexwei.org (self-authored) lists 'D.E. Shaw & Co., Quantitative Research Intern, Summer 2019' under Internships, i.e. an actual paid internship, not a competition, camp, insight day, or offer. Identity is supported: the same CV matches the supplied anchors (IOI 2015 gold for Team USA, Harvard A.B./S.M., UC Berkeley PhD, OpenAI MTS), and D.E. Shaw & Co. (deshaw.com) is a quantitative hedge fund, so the employer genuinely operates in quantitative finance. Only this one quant-finance role is documented; his other internships (Google SWE 2017, Microsoft Research 2020, Meta AI/FAIR 2022) are technology/AI, not trading.Nov 25, 2024
• Identity confirmed: CV at alexwei.org lists IOI Gold Medal 2015, Harvard 2016-2020, UC Berkeley PhD 2020-2023, and OpenAI MTS from Jan 2024 - matching the IOI 2015 Team USA, LinkedIn ayw31, and alexwei.org anchors supplied.Nov 25, 2024
• OpenAI, Member of Technical Staff, from January 2024 (listed as current role on his own CV).Nov 25, 2024
• D.E. Shaw & Co., Quantitative Research Intern, Summer 2019 - a real paid quantitative-research internship, distinct from competitions or insight programs.Nov 25, 2024
• D.E. Shaw & Co. is a New York-based quantitative hedge fund; this classifies the summer-2019 internship as a genuine quantitative-finance role rather than a name-only match.Nov 25, 2024
• IOI 2015 Team USA gold medalist (100 pts, 18/322 worldwide, 94.72%).
• Education: Phillips Exeter Academy diploma 2012-2016; Harvard A.B. summa cum laude in CS and Mathematics 2016-2020 plus S.M. in CS 2019-2020; UC Berkeley Ph.D. in CS 2020-2023 advised by Nika Haghtalab, Michael I. Jordan, and Jacob Steinhardt.
• Additional internships: Google Software Engineering Intern Summer 2017; Microsoft Research Undergraduate Research Intern Summer 2020; Meta AI (FAIR) Research Intern Summer 2022.Nov 25, 2024
• Lead author of 'Jailbroken: How Does LLM Safety Training Fail?' (NeurIPS 2023 Oral) and co-author of the Diplomacy human-level play paper in Science (2022).2023
• Awards: Meta Research PhD Fellowship (Economics & Computation) 2022-2023; NSF Graduate Research Fellowship 2020-2023; SODA 2019 Best Student Paper; INFORMS Auctions & Market Design Rothkopf Prize 3rd place; NeurIPS 2023 Oral.
• Public X account @alexwei_ describes him as 'Reasoning @OpenAI @UCBerkeley PhD '23 | @Harvard '20'.