CV

Jiayi Guo

• Email: richarf22@foxmail.com
• Research Interests: LLM evaluation, RAG, diffusion models, AutoML, and trustworthy AI systems
• Download PDF CV

Education

University of California, San Diego — M.S. in Computer Science Sep. 2026 – Jun. 2028

University of Electronic Science and Technology of China & University of Glasgow — B.Eng. in Communication Engineering Sep. 2022 – Jul. 2026 · GPA: 3.72/4.00

Publications

• IEEE BIBM 2025 — J. Guo, L. Zhang, and Y. Shen. A Human-Centered Multimodal AutoML Framework for Discriminative and Generative Tasks with Large Language Models.
• Submitted to EMNLP 2026 — L. Zhang and J. Guo. When Do LLM Agents Treat Surface Noise Differently from Semantic Noise? A 68-Cell Measurement Study with a Held-Out Trace-Level Validation.

Research Experience

Johns Hopkins University — Large Language Models for Automated Machine Learning (Feb. 2024 – May 2025)

• Led an independent project applying LLMs to discriminative and generative ML workflows across multiple datasets and task settings.
• Evaluated the framework and conducted a 25-participant user study; wrote the paper accepted by IEEE BIBM 2025.

National Supercomputing Center — Diffusion Models for Virtual Try-On Systems (Sep. 2024 – Apr. 2025)

• Developed a diffusion-based virtual try-on framework for single-image garment replacement, including staged and partial replacement for sleeve changes.
• Improved visual realism and garment-fitting quality through model and pipeline adjustments.

UESTC — Generative AI in Semantic Communication Systems (Apr. 2024 – May. 2025)

• Explored latent-space compression and reconstruction, including underwater signal recovery settings.
• Adapted AI models to complex-valued communication scenarios.

Projects

• Resource-Aware LLM/RAG Evaluation Service — Built a local-first FastAPI service with document retrieval, a JSON benchmark schema, quality and reliability metrics, optional DeepSeek judging, and Markdown/HTML reports.
• Kaggle Playground Series: Predicting Irrigation Need — Built a tabular ML pipeline and ranked 16th of 4,315 participants (top 3%).
• Advertiser Growth CRM Backend Extension — Extended Spring PetClinic REST with advertiser-account workflows, RBAC, optimistic locking, transactional outbox events, Kafka retry handling, Redis caching, health checks, and Prometheus metrics.
• Trustworthy Code-Change Review Agent — Built a typed multi-agent review service with FastAPI/CLI interfaces, evaluation and observability, prompt-injection defenses, PII/credential redaction, and 27 unit/integration tests.

Skills

• Languages: Python, C, MATLAB
• ML/Data: PyTorch, NumPy, SciPy, Pandas, Matplotlib, tabular ML, model evaluation
• Engineering: FastAPI, Docker, REST APIs, JSON, Git, LaTeX
• AI Areas: LLM evaluation, RAG, diffusion models, AutoML