CV
Education
- Ph.D in Computer Science, Louisiana State University, Baton Rouge, USA, 2021-2026 (expected)
- B.S. in Information and Computation Science(Math), Northeastern University, Shenyang, CN, 2016-2020
Work experience
- May 2025 - present: Machine Learning Engineer
- Louisiana State University
- Designed a LLM multi-agent framework for athlete injury analysis, rehabitation, and coversation interface with users (coach, athletes)
- Supervisor: Dr.Mingxuan Sun, Nicholas P Totaro
- Aug 2021 - present: Research Assistant
- Louisiana State University
- Designed a spatiotemporal fault-injection framework (STAFI) for ADAS, integrating Progressive Metric-guided Bit Search and context-aware timing to reveal safety-critical vulnerabilities in OpenPilot–CARLA closed-loop driving simulations.
- Built multi-level text-aligned time-series forecasting models that decompose signals into trend/seasonal/residual components and align them with interpretable language anchors to enhance LLM reasoning and forecasting accuracy.
- Proposed a practical backdoor attack that swaps in a malicious attention head, achieving high attack success with minimal clean-accuracy impact and strong resilience to modern defenses.
- Supervisor: Dr.Mingxuan Sun
- Summer 2021: Research Assistant
- Northeastern University
- Implemented and optimized semantic segmentation models for industrial predictive-maintenance applications in collaboration with Ansteel Group.
- Reduced system downtime by 30% through image-based defect detection and automated data labeling pipelines.
- Supervisor: Dr.Zhuofu Deng
Skills
- Language: Python, C++, Java, MATLAB
- Machine learning: pytorch, tensorflow, Huggingface, Scikit-learn, ONNX Runtime, OpenCV
- Tools: Numpy, Pandas, Docker, AWS, Git, Jupyter, Linux
Publications
Talks
Enhancing Time Series Forecasting via Multi-Level Text Alignment with LLMs
Talk at LSU CSE Student Research Lagniappe, Baton Rouge, LA, USA