Muhammad Faaiz Taufiq

Muhammad Faaiz Taufiq is a research scientist at ByteDance Seed in London working on LLM reasoning and multi-agent collaboration. He holds a PhD in machine learning from the University of Oxford (2020-2024), supervised by Rob Cornish, Arnaud Doucet and Yee Whye Teh and funded by Google DeepMind, with prior study at the University of Cambridge. As a school student in Karachi he represented Pakistan at the 2013 International Mathematical Olympiad, earning an Honourable Mention. His published work spans conformal prediction and off-policy evaluation, causal inference, interpretability and algorithmic fairness, appearing at NeurIPS, ICML and AISTATS. Early in his career he held a quantitative placement on Citi's Exotics Trading Desk and technology roles at Morgan Stanley and J.P. Morgan.

Notes8

• Quant employment verification: CONFIRMED. The LinkedIn profile index lists a quantitative-finance role: Quantitative Placement Analyst at Citi, London, Jun-Aug 2018, on the Exotics Trading Desk, where he created a Risk Analytics Tool for daily risk-metric calculations across trading sub-desks (LinkedIn index). Citi is a global bank and its exotics/derivatives trading desk is a quantitative-finance function, so this is a genuine placement/internship on a trading desk, not a competition, insight day, camp or incoming offer. Identity is supported: the same LinkedIn profile is cross-linked from his personal site and matches his Oxford PhD and ByteDance Seed role. Caveat: the Citi record rests on the LinkedIn profile as surfaced by the index; a direct fetch of the public LinkedIn page did not expose the experience section, so the role is index-sourced rather than independently corroborated by a second primary source. Other finance-adjacent roles are technology, not quant trading: Morgan Stanley Technology Analyst/Associate (2018-2020) and J.P. Morgan Summer Technology Analyst (2017, Cyber Security team).2018
• Completed a PhD in machine learning at Oxford (2020-2024) supervised by Rob Cornish, Arnaud Doucet and Yee Whye Teh, funded by Google DeepMind; supervisors are named on his own site and the Oxford Statistics student list.2024
• Received an Honourable Mention at IMO 2013 for Pakistan (score 8, rank 355); his LinkedIn education places him at Karachi Grammar School in that period, tying the olympiad identity to the same person.2013
• Represented Pakistan at the 2013 International Mathematical Olympiad while at Karachi Grammar School, scoring 32.8%; his LinkedIn education lists Karachi Grammar School 2012-2014.2013
• Research Scientist at ByteDance Seed, London, working on LLM reasoning and multi-agent collaboration, per his self-authored personal site bio.
• Published PhD thesis 'Uncertainty Quantification and Causal Considerations for Off-Policy Decision Making' (arXiv:2502.06011).2025
• Career arc: STEM internships (Dyson R&D 2016, J.P. Morgan cyber 2017, Citi exotics-trading desk summer 2018) into bank technology roles (Morgan Stanley 2018-2020), then a pivot to ML research (Oxford PhD 2020-2024, AWS causality intern 2022, ByteDance 2023-present); the quantitative-finance stint is a single summer placement and the only trading-desk role in the timeline.2024
• Google Scholar lists affiliation 'ByteDance Seed' with a verified bytedance.com email, an independent signal corroborating his current employer; recurring co-author Jean-Francois Ton is also at ByteDance Seed.
• Public contact email listed on his personal site: faaiz.taufiq (at) bytedance.com.
• Organising-committee member for the ELLIS RobustML workshops (two robustness workshops organised in 2024), per his personal site.2024
• PhD supervisors Rob Cornish, Arnaud Doucet and Yee Whye Teh are named both on his site and on the Oxford Statistics doctoral-students page, cross-confirming the Oxford DPhil.2023
• Named in a 2014 Pakistani press report as one of two students granted a Cambridge 800th Anniversary Scholarship for undergraduate study - a scholarship, not employment.May 2014