Yiang Zhou
About
I'm a Computer Science student at Oregon State University. My internships have centered on AI and backend engineering: RAG pipelines and multi-agent workflows with LangChain and LangGraph, high-concurrency FastAPI services, an AI agent platform for automated telescope scheduling and surveys, and the development and testing of AI algorithms. In research, I've worked on parametric, interpretable models of human vibroacoustic sounds (paper published at IEEE EMBC 2026) and differential privacy for federated CNN training.
Education
Oregon State University
2026 – Present · Oregon, United States
B.S. in Computer Science
Oregon, United States
Universiti Sains Malaysia (USM)
2023 – 2026 · Penang, Malaysia
Computer Science coursework
Penang, Malaysia
Research, engineering & AI agents
Closed source
Bridge
ContributorBuilt real-time voice input for the macOS app: live speech-to-text typed into other apps.
AFK AI
Engineering
High-concurrency APIs
Async FastAPI services at 200–300 QPS; hybrid Milvus + Elasticsearch retrieval with RRF re-ranking improved knowledge-base Q&A accuracy by 10–25%.
Chongqing AI Institute, SJTU
AI agents
RAG & multi-agent workflows
Self-correcting retrieve–generate–verify workflows with LangChain and LangGraph, for medical knowledge Q&A.
Langlands Technology
Work Experience
AFK AI, INC.AI / Executive InternMar 2026 – Sep 2026HybridMar 2026 – Sep 2026Hybrid
- Algorithms & engineering: Assisted in the development and testing of AI algorithms and supported engineering workflows.
- Data & operations: Supported data processing, analysis, and cross-functional execution across technical and business tasks.
- Coordination: Participated in team meetings, project coordination, and discussions on product and engineering initiatives.
- Industry events: Represented AFK AI at NVIDIA GTC 2026 (March 2026).
Shanghai Astronomical Observatory, Chinese Academy of SciencesResearch InternAug 2025 – Oct 2025Remote · Part-timeAug 2025 – Oct 2025Remote · Part-time
- Full-stack AI agent platform: Independently built an end-to-end AI agent research platform for automated telescope scheduling and cosmological surveys.
- System design: Designed the interactive observation UI, the Python backend logic for scheduling telescope observations, and the database schema for observation parameters and astronomical images.
- DevOps: Managed server deployment, environment setup, SSL certificates, and domain setup.
- AI agents
- Python
- Database design
- Server deployment
- SSL certificates & domains
Langlands TechnologySoftware Engineering InternAug 2025 – Sep 2025Suzhou, ChinaAug 2025 – Sep 2025Suzhou, China
- Core RAG algorithms: Developed RAG pipelines and multi-agent workflows with LangChain and LangGraph, including a self-correcting retrieve–generate–verify mechanism, and built medical knowledge Q&A with ChromaDB and DeepSeek.
- Cloud infrastructure: Deployed Alibaba Cloud RDS and built a hybrid cloud–edge architecture with local HuggingFace embedding models.
- Data engineering: Implemented Pandas ETL scripts for multi-format medical documents (PDF OCR, Excel, CSV), improving backend reliability.
- LangChain
- LangGraph
- ChromaDB
- DeepSeek
- HuggingFace embeddings
- Alibaba Cloud RDS
- Pandas
Chongqing AI Institute, Shanghai Jiao Tong UniversityPython Backend / AI InternSep 2024 – Oct 2024Chongqing, ChinaSep 2024 – Oct 2024Chongqing, China
- RAG optimization: Enhanced the RAG system with dual-path Milvus + Elasticsearch retrieval and RRF re-ranking, improving knowledge-base Q&A accuracy by 10–25%.
- High concurrency: Built high-concurrency FastAPI services for a model demo platform; async refactoring (run_in_threadpool) resolved blocking at 200–300 QPS.
- API delivery: Delivered Pydantic models, SSE streaming, and automated API testing with Apifox.
- Python
- FastAPI
- Milvus
- Elasticsearch
- Pydantic
- SSE streaming
- Apifox API testing
Open Source
- GitHub
Comma
Contributor
Research Experience
Aug 2025 – Oct 2025Remote · Part-time
Université de Bourgogne
Department of Applied Mathematics
Research Intern
- Project: Open-source Simulator for Human Vibroacoustic Sounds.
- Core work: Developed parametric and interpretable vibroacoustic models (envelope and temporal modeling, spectral analysis, signal generation) and reproducible data pipelines to address the scarcity of medical auscultation data.
- Outcomes: Paper published at IEEE EMBC 2026 and presented at the conference in July. Invited speaker at GOSIM 2025 in Hangzhou.
- Envelope & temporal modeling
- Spectral analysis
- Signal generation
Jul 2025Shanghai, China
Sino-French AI Summer School, Shanghai Jiao Tong University
Research Member
- Project: Privacy protection for federated convolutional neural networks with differential privacy.
- Implementation: Implemented differential privacy mechanisms for federated CNN training and designed the noise injection and privacy budget control strategies.
- Outcome: The final project received the Innovation Achievement Award; paper in progress.
- Federated learning
- Differential privacy
- CNN
Publications, Talks & Awards
- Publication
A Dynamic Parametric Simulator for Fetal Heart Sounds
Yingtong Zhou, Yiang Zhou, Zhengxian Qu, Kang Liu, Ting Tan
IEEE EMBC 2026Presented at the conferenceJul 2026
- Invited talk
Invited speaker at GOSIM 2025, Hangzhou
Human vibroacoustic simulatorUniversité de Bourgogne
- Award
Innovation Achievement Award
Final projectSino-French AI Summer School, Shanghai Jiao Tong UniversityJul 2025
Skills
AI & LLM applications
- RAG
- LangChain
- LangGraph
- Multi-agent workflows
- AI agents
- DeepSeek
- HuggingFace embeddings
- RRF re-ranking
Backend & APIs
- Python
- FastAPI
- Pydantic
- SSE streaming
- Apifox API testing
Data & retrieval
- Milvus
- Elasticsearch
- ChromaDB
- Pandas
- ETL
- PDF OCR
- Database design
Cloud & DevOps
- Alibaba Cloud RDS
- Cloud–edge architecture
- Server deployment
- SSL certificates & domains
ML & signal research
- Federated learning
- Differential privacy
- CNN
- Spectral analysis
- Signal generation
- Envelope & temporal modeling
