cv

Machine Learning Engineer Intern

Epoching AI

Developed an end-to-end watermark removal service and optimized training pipelines for e-commerce imagery.

• Fine-tuned pre-trained ViT model, achieving 30% accuracy improvement.
• Deployed scalable FastAPI/Docker service with Prometheus monitoring.

Machine Learning Engineer Intern

Ytell Network Technology Co., Ltd.

Designed and implemented multimodal knowledge bots and automated pipelines for order processing.

• Integrated VLMs with backend, improving response accuracy by 10%.
• Developed automated ETL workflows processing 1,000+ daily requests with 95% accuracy.

Master of Science

Carnegie Mellon University

Artificial Intelligence Engineering - Information Security

• Large Language Models
• AI Systems and Tool Chains
• Computer Systems

Bachelor of Engineering

University of International Business and Economics

Data Science and Big Data Technology

• Machine Learning
• Data Structure and Algorithm
• Natural Language Processing
• Big Data Analysis

Memory Allocator

Implemented a high-performance dynamic memory allocator in C with tools for debugging and optimization.

• Achieved 75%+ memory utilization through boundary tag coalescing.
• Integrated GDB debugging tools for block alignment validation.

Transformer Decoder

Built a decoder-only Transformer with custom tokenizer and training pipeline.

• Published trained model to HuggingFace Hub.
• Integrated Weights & Biases for experiment tracking.

Self-LLM: Open-Source LLMs Deployment Guide

Collaborated on deployment guides for GLM4, enabling developers to build advanced LLM applications.

• 10k+ GitHub stars for deployment guides.

Keep Knowledge in Perception: Zero-shot Image Aesthetic Assessment

Developed a novel zero-shot framework for fine-grained aesthetic assessment, outperforming SOTA approaches.

• ICASSP 2024 Oral Presentation.
• Created a large-scale aesthetic dataset with 175k expert critiques.