CV

Research Assistant

KAUST, Prof. Basem Shihada

• Research on AI Scientist, Self-Improving AI, and Multi-Agent RL
• Exploring generative models for scientific discovery
• Developing multi-agent systems for collaborative AI research

Research Intern

MoleculeMind

• Designed and implemented an unconditional protein backbone generation model using SE(3) diffusion (DDPM on SO(3))
• Developed an epitope-conditioned binder design framework via cross-attention over target structure features
• Extended the SE3Diffusion Binder module to support multi-chain protein binder generation
• Investigated Flow Matching on SE(3) as an alternative to diffusion, achieving faster convergence
• Systematically analyzed iPAE and distogram metrics; deployed DiffAffinity online service with JAX compatibility

Research Assistant

Simon Fraser University / Natera, Prof. Wuyang Chen & Dr. Zijun Zhang

• Developed TEACUP framework for training-free evaluation of clinical AI systems
• Achieved 90% reduction in computational costs
• Implemented ensemble modeling for clinical decision-making

Research Assistant

Northwestern University, Prof. Han Liu

• Designed evaluation metrics for GenomeOcean DNA sequence generation model
• Conducted comparative analysis of ORF length distributions
• Analyzed codon bias patterns for biological realism assessment

Research Assistant

UC Berkeley, Dr. Wuyang Chen

• Developed training-free data augmentation design principles
• Published findings in ICML 2024 Workshop AI4Science

MS/PhD

King Abdullah University of Science and Technology (KAUST)

Computer Science

• Member of the Intelligent Systems Lab
• Advisor - Prof. Basem Shihada
• Research Focus - AI Scientist, Self-Improving AI, and Multi-Agent RL
• Relevant Coursework - Deep Learning for Visual Computing, Scientific Visualization, Robotics

B.S.

University of Science and Technology of China (USTC)

Life Sciences (Biotechnology)

• Department of Life Sciences and Medicine
• Core Courses - Linear Algebra B1, Electromagnetism B, Undergraduate Research Project (A+)
• Relevant Coursework - Molecular Biology, Biochemistry, Structural Biology, Bioinformatics, Probability & Statistics
• Interdisciplinary background bridging wet-lab biology with computational modeling for AI-driven research

D-DOIT: Training-free Adaptation of Discrete Diffusion via Doob's h-Transform

First-author paper on training-free adaptation of discrete diffusion models using Doob’s h-transform.

Training-free Design of Augmentations with Data-centric Principles

First-author workshop paper on training-free data augmentation design.

Annealed Co-Generation: Disentangling Variables via Progressive Pairwise Modeling

AnchorGen: Anchored Optimization for Customizable Generative 3D Design

GenomeOcean: Efficient Foundation Model for Genome Generation

Training-free Design of Deep Networks as Ensembles of Clinical Experts

Outstanding Graduate of USTC, Class of 2025

USTC

University-level Outstanding Project (Undergraduate Innovation and Entrepreneurship Training Program)

USTC

University-level Outstanding Project (College Student Research Program)

USTC

8th National College Student Life Science Competition, First Prize

National Level

Outstanding Student Scholarship

USTC

Outstanding Student Scholarship

USTC

Outstanding Student Scholarship

USTC

Outstanding Student Scholarship

USTC