Yan (Melody) Zhao
📧 zhao.y4@northeastern.edu • 📱 (206) 843-7025 • LinkedIn • GitHub • Hugging Face • Seattle, WA
🎓 Education
Northeastern University • Seattle, WA
Master of Science in Computer Engineering • GPA: 4.0/4.0
Jan 2025 - May 2027
🛠️ Technical Skills
- Languages & AI: Python, Java, TypeScript • PyTorch, TensorFlow, JAX
- Frameworks: LangChain, Hugging Face, RAG, VLLM, ChromaDB
- Cloud & DevOps: AWS, Google Cloud, Docker, Kubernetes, CI/CD
- Certifications: AWS Solutions Architect, Google Cloud ML Engineer
💼 Professional Experience
Human Ageing Genomic Resources • May 2025 - Sep 2025
AI Engineer Intern • Seattle, WA
- Led development of LLM-Powered Biomedical Chatbot
- Built end-to-end RAG system with Python, ChromaDB, and Streamlit
- Reduced researcher query time by 60% through intelligent routing
- Implemented real-time streaming, citation tracking, and history management
- Engineered robust data pipelines and fine-tuned Gemma LLM
- Unified diverse biomedical datasets and optimized vector embeddings
- Achieved efficient training using LoRA (<1% parameters) on Cloud TPU
- Deployed high-performance vLLM inference with sub-second responses
</div>
### Tianjin Motor Dies Co., Ltd. Sep 2008 - Dec 2018
**Technical Solutions Engineer & Project Manager** Tianjin, China
- Led **$10M+** automotive projects for Fortune 500 clients (Tesla, Ford, GM)
- Managed global engineering teams of 50+ professionals
- Improved delivery efficiency by **30%** through data-driven workflows
### Northeastern University Sep 2025 - Present
**Teaching Assistant** Seattle, WA
- Lead weekly Python programming sessions for 30+ students (INFO5002)
- Create learning materials and provide one-on-one mentoring
🚀 Featured Projects
Deep Learning Fundamentals & Autograd Engine
- Built autograd engine and MLP training loop from scratch
- Implemented backpropagation with topological sort
- Tech: Python, Neural Networks, PyTorch
High-Performance Distributed ML
- Implemented JAX data parallelism and ZeRO optimization
- Analyzed VLLM inference performance optimization
- Tech: JAX, PyTorch, VLLM
Multi-Channel E-commerce Platform
- Built layered architecture with analytics dashboard
- Tech: Java, Spring Boot, React
- Built end-to-end RAG system with Python, ChromaDB, and Streamlit
- Reduced researcher query time by 60% through intelligent routing
- Fine-tuned Gemma LLM using LoRA, achieved sub-second inference
- Engineered data pipelines for biomedical datasets
Tianjin Motor Dies Co., Ltd. • Sep 2008 - Dec 2018
Technical Solutions Engineer & Project Manager • Tianjin, China
- Led $10M+ automotive projects for Fortune 500 clients (Tesla, Ford, GM)
- Managed global engineering teams of 50+ professionals
- Improved delivery efficiency by 30% through data-driven workflows
Northeastern University • Sep 2025 - Present
Teaching Assistant • Seattle, WA
- Lead weekly Python programming sessions for 30+ students (INFO5002)
- Create learning materials and provide one-on-one mentoring
🚀 Featured Projects
Deep Learning Fundamentals & Autograd Engine
- Built autograd engine from scratch with backpropagation
- Implemented MLP training loop with gradient descent
- Tech: Python, Neural Networks, Backpropagation
High-Performance Distributed ML
- Implemented JAX data parallelism and ZeRO optimization
- Analyzed VLLM inference performance and KV cache
- Tech: JAX, PyTorch, VLLM
Multi-Channel E-commerce Platform
- Built layered architecture with analytics dashboard
- Tech: Java, Spring Boot, React
🌟 Open Source Contributions
- JAX: Performance optimizations and API improvements
- Hugging Face: Model implementations and documentation
- LangChain: RAG architecture enhancements
Last updated: January 2025