Shayan Pardis

Shayan Pardis is an ML research engineer at Cursor and a recent MIT graduate in Computer Science and Mathematics. He represented Iran at the International Olympiad in Informatics 2020 (gold, rank 10) and won gold at Iran's national informatics olympiad in 2019 before moving from Sharif University of Technology to MIT in 2022. His research spans reinforcement learning, multi-agent learning, and robotics, including a first-author IROS 2024 paper on probabilistic homotopy optimization for humanoid motion planning and MIT CSAIL work on diffusion-based planning. He interned at Citadel as a Quantitative Developer in Central Risk Engineering (summer 2024), where he worked on distributed infrastructure rather than trading strategy.

Notes8

• Quant employment verification: CONFIRMED. His CV lists a real internship - 'Citadel LLC, Quantitative Developer Intern at Central Risk Engineering, Jun 2024 - Aug 2024' - and Citadel is a major quantitative hedge fund/market maker; the identity is supported (name + MIT + IOI 2020 gold + matching personal site/GitHub/LinkedIn). The role is quantitative-developer infrastructure work inside a quant firm, i.e. an actual internship rather than a student program or insight day; the CV also claims a return offer. No other listed employer operates in quantitative finance: Cursor (AI coding tool), Backflip AI (3D CAD), MIT CSAIL and Biomimetic Robotics Lab (academic), Google Summer of Code / QuantumClifford.jl (open source), SIMCON (geometric algorithm design), Carriot (geocoding), Abarkelas (web dev).Aug 2024
• Citadel LLC role is an actual internship, not a student program: Quantitative Developer Intern, Central Risk Engineering, Jun-Aug 2024, New York City; CV describes distributed-system infrastructure work (Kubernetes, gRPC, Cloud Run, Redis) and a parallel testing framework, and states a return offer was secured.Jun 2024
• Career is research/engineering rather than trading: every listed role is software/ML research engineering, with Citadel the only finance-sector role and it titled 'Quantitative Developer' (risk-engineering infrastructure, not a trading or strategy seat).2025
• Identity confirmed: Shayan Pardis, MIT CS (6-3) and Mathematics (18); his personal CV lists IOI 2020 gold (rank 10) for Iran, Iran National Olympiad in Informatics 2019 gold (rank 1), and APIO 2020 silver (rank 24). The caller-supplied Olympiad identity (Shayan Pardis, Iran, IOI 2020) matches this person on the CV's own text.2024
• Cursor role is current as of the CV (ML Research Engineer, Jun 2025-Present, San Francisco), and the LinkedIn index renders the same tenure as '2025-2026' with the shorter title 'Research Engineer' - so the index truncates a role the primary CV states as ongoing; treat the CV as the more current source.Jun 2025
• Public contact email is [contact omitted], published on his personal CV along with links to github.com/Shayan-P, linkedin.com/in/shayan-pardis, and his Google Scholar profile.2025
• Synthesis: a research-track trajectory, not a finance track - the single finance-sector role (Citadel, summer 2024) sits between two AI/ML research roles (Backflip Jan 2025, Cursor Jun 2025), and every other entry is academic research, open source, or engineering, which reads as a computational-research career that briefly intersected quantitative finance rather than a quant career.2025
• Identity is triple-anchored across independent surfaces: the CV (shayanp.me/cv.pdf), GitHub profile github.com/Shayan-P, and LinkedIn /in/shayan-pardis all resolve to the same MIT student with matching name, email ([contact omitted]), and links, making the profile URL claims mutually corroborating rather than index-only.2025
• Published first-author research: 'Probabilistic Homotopy Optimization for Dynamic Motion Planning' (arXiv 2408.12490, IROS 2024), from MIT Biomimetic Robotics Lab work with Matthew Chignoli under Prof. Sangbae Kim, plus a 4x QP-controller speedup for the MIT Humanoid.2024
• Research focus stated on his own site: reinforcement learning, multi-agent learning, and meta-learning for robotics, aiming at a 'society of AI agents'; MIT work with the Learning in Intelligent Systems group (Leslie Kaelbling, Yilun Du) on diffusion-based planning.2025
• Migrated from Sharif University of Technology (Tehran) to MIT after two years (Sharif Sep 2020-May 2022, GPA 3.9/4.0), which places his move to the US in 2022 and explains the MIT research roles beginning Feb 2023.May 2022
• Backflip AI internship was one month (Jan 2025 - Jan 2025) training Vision Transformers to convert sketches into parameterized shapes for a 3D CAD editing tool - a short research stint, distinct from his ongoing Cursor tenure.Jan 2025