CV

Full CV (PDF) One-page résumé (PDF)

Education

• B.Sc. in Artificial Intelligence, MBZUAI, Abu Dhabi — Aug 2025 – May 2029 (expected). GPA 3.8/4.0.
  • Honors: Dean’s List (inaugural cohort, Fall 2025, top 10%); Sheikh Tahnoon bin Zayed Scholarship in AI Excellence; Pioneer Scholarship.
  • Courses: Applied Machine Learning (ML7501, M.S. course), Algorithms & Data Structures, Probability & Statistics, Calculus & Linear Algebra, Introduction to AI, Python Programming, Software, Web & Mobile Engineering, Physics & Life Science in the Age of AI.
• High school, Beijing National Day School, Beijing — Sep 2019 – Jul 2025. GPA 4.2/4.3 (unweighted).
  • Advanced Placement program jointly with Wasatch Academy.
  • Courses: Calculus III, Differential Equations, Linear Algebra, Probability Theory, Mathematical Modeling, Sociology, Marine Biology.
  • SAT 1550 (Math 800, English 750); TOEFL 111/120; CEFR C1 (191/200); 11 AP exams with eight 5s (Calculus BC, Physics C: Mechanics, Physics 1, Chemistry, Biology, Computer Science A, Statistics, Psychology).

Experience

• Cyber-Physical Intelligence Lab & AAAAA Community, Researcher — Oct 2025 – present
  • AI labs across MBZUAI, McGill University & Mila. CPIL advances trustworthy, efficient, and agentic AI for the cyber-physical world; AAAAA Community is a research collective spanning universities, labs, and industry. Directly mentored by lab founder Prof. Steve Liu and senior member Ye Yuan.
  • SPADE: Support-Proximity Augmented Diffusion Estimation for Offline Black-Box Optimization (co-first author, Oct 2025 – Jan 2026): experimented with conditional diffusion models as calibrated surrogates for offline optimization, with a kNN support-proximity prior that keeps the optimizer out of unsupported regions. Accepted at ICML 2026 (poster presented in Seoul) and the ICLR 2026 DeLTa workshop. Code · Project page · Video · Poster · Slides
  • Surrogate-Guided Memory Retrieval for Agents (co-first author, Apr 2026 – present): proposed and benchmarked surrogate-guided offline training of agent memory retrieval, scoring each memory by how much it helps downstream tasks. Submitted to the NeurIPS 2026 BiAlign workshop.
  • QUACK: Questioning, Understanding, and Auditing Communicated Knowledge in Multimodal Social Deduction Agents (Feb 2026 – present): audited whether agents’ claims in social deduction games are grounded in what they saw and did. Accepted at EMNLP 2026 (Main Conference). Code · Dataset
  • Discrete Diffusion Models: A Unified Framework from Tokenization to Generation (Mar 2026 – present): investigated AI-for-Science applications for a survey unifying discrete diffusion in one design space. Under review at TMLR. Paper list
  • Herculean: An Agentic Benchmark for Financial Intelligence (Mar – May 2026): tested agents on representative financial intelligence workflows — trading, hedging, market insights, and auditing. Accepted at ACL 2027. Code · Dataset
• Precision-medicine AI Lab, Researcher — Mar – May 2026
  • AI-for-Science lab across MBZUAI & Harvard Medical School, integrating multi-scale biomedical data for precision medicine. Directly mentored by lab founder Prof. Jun Wen.
  • Semi-supervised High-order Relation Learning: trained a multi-head-attention (MHA)-based model on a comprehensive drug combination–disease relationship dataset.
• Paper Reviewer — Feb 2026 – present
  • Reviewed for top ML venues: the ICLR 2027 main conference and the ICLR 2026 Workshop on Foundation Models for Science (FM4Science).
• GenBio AI, Research Engineer Intern — Mar 2026 – present
  • AI-for-Science startup co-founded by world-renowned scientists including Nobel laureate Prof. David Baker, building AIDO, an AI-Driven Digital Organism. Supervised by CTO Prof. Le Song.
  • BioXArena: Benchmarking LLM Agents on Multi-Modal Biomedical Machine Learning Tasks: designed multi-modal biomedical ML agent tasks and benchmarked agents (76 end-to-end tasks across 9 domains); key contributions to the paper’s formulation. Accepted at NeurIPS 2026 (Evaluations & Datasets Track). Code · Project page · Dataset
  • VCHarness: optimized the efficiency of the MCTS-based harness of GenBio’s autonomous system for building virtual-cell models, and benchmarked it on BioXArena.
• Shanghai Academy of AI for Science, Research Engineer Intern — May 2026 – present
  • AI-for-Science R&D institute focused on industrial applications. Engineered the A2A (Agent2Agent) protocol and researched generative models.
  • Decentralized Agentic Reasoning: optimized a decentralized agent swarm built on the established A2A protocol.

Publications

Discrete Diffusion Models: A Unified Framework from Tokenization to Generation

Ye Yuan, Weien Li, Rui Song, et al. (including Yonghan Yang), Dawn Song, Philip S. Yu, and Xue Liu. (2026). "Discrete Diffusion Models: A Unified Framework from Tokenization to Generation." arXiv:2607.13431. Under review at TMLR.

QUACK: Questioning, Understanding, and Auditing Communicated Knowledge in Multimodal Social Deduction Agents

Ye Yuan, Rui Song, Weien Li, et al. (including Yonghan Yang), and Xue Liu. (2026). "QUACK: Questioning, Understanding, and Auditing Communicated Knowledge in Multimodal Social Deduction Agents." EMNLP 2026. arXiv:2605.27068.

Herculean: An Agentic Benchmark for Financial Intelligence

Xueqing Peng, Zhuohan Xie, Yupeng Cao, et al. (including Yonghan Yang). (2026). "Herculean: An Agentic Benchmark for Financial Intelligence." ACL 2027. arXiv:2605.14355.

BioXArena: Benchmarking LLM Agents on Multi-Modal Biomedical Machine Learning Tasks

Loka Li, Duzhen Zhang, Xingbo Du, et al. (including Yonghan Yang), Bin Zhang, and Le Song. (2026). "BioXArena: Benchmarking LLM Agents on Multi-Modal Biomedical Machine Learning Tasks." NeurIPS 2026. arXiv:2605.15766.

Support-Proximity Augmented Diffusion Estimation for Offline Black-Box Optimization

Yonghan Yang*, Ye Yuan*, Zipeng Sun, Linfeng Du, Bowei He, Haolun Wu, Can Chen, and Xue Liu. (2026). "Support-Proximity Augmented Diffusion Estimation for Offline Black-Box Optimization." Proceedings of the 43rd International Conference on Machine Learning (ICML 2026), PMLR 306, Seoul, South Korea. (* equal contribution)

Surrogate-Guided Memory Retrieval for Autonomous Agents

Yonghan Yang*, Ye Yuan*, et al. Surrogate-Guided Memory Retrieval for Autonomous Agents. Submitted to the NeurIPS 2026 BiAlign Workshop, 2026. (* equal contribution)

Semi-Supervised High-Order Relation Learning for Drug–Combination–Disease Prediction

Yonghan Yang, et al. Semi-Supervised High-Order Relation Learning for Drug–Combination–Disease Prediction. Manuscript in preparation, 2026.

Awards and honors

• MBZUAI Dean’s List, inaugural cohort (Fall 2025, top 10%) · Sheikh Tahnoon bin Zayed Scholarship in AI Excellence & Pioneer Scholarship (2025)
• 2025 IMMC — Outstanding (top 1%, Greater China round) & Meritorious (top 10%, International round)
• iGEM high-school section — Gold Medal & Top 10 for three years (2023–2025); 2025 Best Measurement & Best Part Collection as instructor; 2024 Best Wiki nominee as student leader
• 2024 John Locke Institute Global Essay Prize — Finalist & Merit (Economics)
• 2024 Goi Peace Foundation International Essay Contest for Young People — Finalist (top 100 of 10,233)
• 2024 ARML — National Silver · 2024 Purple Comet! Math Meet — China 2nd, global 16th · 2024 Johns Hopkins Math Tournament — Team Silver, Power & Relay Round Bronze
• 2023 British Biology Olympiad — Global Gold · 2022 UKBC Intermediate Biology Olympiad — Bronze
• 2023 COMAP HiMCM — Honorable Mention · 2022 Junior Canadian Chemistry Olympiad — China Gold, global Bronze · 2022 Berkeley Math Tournament — Honorable Mention
• 2022 China Thinks Big — research paper accepted to the Digital Library collection
• AP Scholar (2023) & AP Scholar with Distinction (2024, 2025) · 2019 FIRST Tech Challenge — Beijing regional first prize

Activities

• Math Modeling Competitions (COMAP, MAA & AMS), team leader — Jan 2025 – present
  • 2025 IMMC Outstanding (Greater China, top 1%) and Meritorious (International, top 10%). Built MILP scheduling models with PuLP and typeset both papers in LaTeX.
• Synthetic Biology, iGEM — BNDS-China (team wiki) — Oct 2022 – present
  • Contributed as instructor (2025–2026), student leader (2024) and member (2023). Top 10 teams and gold medal in 2023–2025, plus Best Measurement and Best Part Collection in 2025. Developed a programmable probiotic platform against gut dysbiosis (2024) and one for feline gut health (2025); built the dry-lab modeling. See the project page.
• Rhino-Bird Science Talent Advancement Program (BNRist, Tsinghua & Tencent) — Apr – Aug 2024
  • 30-hour research training; implemented the attention mechanism in PyTorch; researched dynamic game theory in college admissions.
• Jingwan Applied Math Summer Camp (Beijing Normal University), team leader — Jul – Aug 2023
  • Modeled honeybee colony collapse disorder from hive data, leading a team of three.
• School clubs, Beijing National Day School — Sep 2022 – Jul 2025
  • President of biology club BioCamp; co-founder of math club MathCorner; vice president of the origami club; school badminton team.
• Volunteering — Sep 2023 – present
  • MBZUAI HCI Symposium 2025 (event operations); IUCN World Conservation Congress 2025 at ADNEC, Abu Dhabi; Beijing environmental-protection volunteer (2024); co-founder of a dyslexia support group.

Skills and interests

• Coding: Python, Rust, TypeScript; deep learning with PyTorch. Writing: Markdown, Typst, LaTeX. Visuals: Manim, PowerPoint, Photoshop.
• AI tooling: Anthropic certificates (Jan 2026) in Building with the Claude API, Claude Code in Action, Model Context Protocol, and AI Fluency; AWS AI Practitioner Learning Plan (Apr 2026).
• Lab: SnapGene, Benchling, AlphaFold, AutoDock Vina, PyMOL.
• Languages: Mandarin (native), English (fluent); learning French and Japanese.
• Sketching, tennis, running (70 km+ a month), fencing, badminton, origami, and oil painting.