CV
Education, research experience, publications, and technical background.
Contact Information
| Name | Ang Qi |
| Professional Title | Applied and Computational Mathematics · USC |
| angqi@usc.edu | |
| Location | Los Angeles, California |
Professional Summary
Undergraduate researcher studying vision-language model robustness, visual reasoning, and targeted model repair.
Experience
-
2025 - present Hybrid
Research Assistant
Physical Superintelligence Lab, University of Southern California
- Researched vision-language model robustness and reliability through failure analysis and targeted post-training.
- Developed pipelines to generate, filter, and evaluate multimodal data across visual reasoning tasks.
- Analyzed failures in object recognition, attributes, spatial relationships, counting, and negation.
- Applied large vision-language models and parameter-efficient fine-tuning to investigate targeted model repair and generalization.
- Contributed to ongoing research on vision-language model robustness, with results being prepared for workshop submission.
-
2023 - 2023 Beijing, China
Research Assistant
Laboratory of Microfabrication, Institute of Physics, Chinese Academy of Sciences
- Assisted with microelectronic substrate fabrication for terahertz communication devices using lift-off processes.
- Operated mask aligners, ICPCVD systems, and electron-beam evaporation tools.
- Supported ultra-low-temperature experiments and quantum transport measurements.
Education
-
2024 - 2027 Los Angeles, CA, U.S.
Bachelor of Science
University of Southern California
Applied and Computational Mathematics
- Expected graduation: May 2027
- Minor: Computer Science
- GPA: 3.76/4.00.
-
2021 - 2024 Beijing, China
High School
Beijing Royal School
- Unweighted GPA: 4.19/4.33 (Top 3%)
- Weighted GPA: 4.74/5.0 (Top 3%)
Publications
-
2026 Large Reward Models: Generalizable Online Robot Reward Generation with Vision-Language Models
IROS 2026 AI Meets Autonomy Workshop
Yanru Wu, Weiduo Yuan, Ang Qi, Vitor Guizilini, Jiageng Mao, Yue Wang
-
2026 FuzzingRL: Reinforcement Fuzz-Testing for Revealing VLM Failures
In Submission
Jiajun Xu, Jiageng Mao, Ang Qi, Weiduo Yuan, Alexander Romanus, Helen Xia, Vitor Campagnolo Guizilini, Yue Wang
Selected Coursework
Mathematics: MATH-226: Calculus III; MATH-225: Linear Algebra and Linear Differential Equations; MATH-407: Probability Theory; MATH-432: Applied Combinatorics; MATH-408: Mathematical Statistics; MATH-430: Number Theory; MATH-458: Numerical Methods (Fall 2026); MATH-446: Data Science with Python (Fall 2026).
Computer Science: CSCI-103: Introduction to C++; CSCI-170: Discrete Methods in Computer Science; ITP-168: Introduction to MATLAB; CSCI-104: Data Structures; CSCI-270: Introduction to Algorithms and Theory of Computing; CSCI-360: Artificial Intelligence: Principles and Foundations; CSCI-467: Foundations of Machine Learning (Fall 2026); TAC-216: Applied Python (Fall 2026).
Skills
Programming Languages: C++, Python, Java, MATLAB, R
Frameworks & Libraries: PyTorch, NumPy
Tools & Methods: Git, Linux, LaTeX, PEFT/LoRA
Awards
-
2026 Academic Achievement Award
University of Southern California
Fall 2025 and Spring 2026.
-
2024 Outstanding Graduate
Beijing Royal School
Scholarship of 50,000 RMB.
-
2022 High Distinction
Australian Science Olympiads Physics
Languages
Mandarin : Native
English : Proficient