本文最后更新于 2025年9月6日 下午
Kaiyuan (Gawain) Tan
AI/ML Engineer
+86 18982160568 | tankaiyuan33@gmail.com | LinkedIn | Portfolio
Professional Summary
AI/ML Engineer specializing in Large Language Models (LLMs) with end-to-end experience in building, fine-tuning (QLora), and deploying scalable AI systems (RAG, Docker). Proven ability to deliver high-precision models and architect robust data pipelines in production environments.
Experience
Sichuan JiXia Tech - Algorithm Engineer Sep 2024 – Present
- Architected an end-to-end document intelligence pipeline using Qwen-VL and PaddleOCR, achieving >93% classification accuracy on complex legal documents.
- Developed and optimized a multi-stage NLP system combining a SpaCy NER model (reaching >95% precision) with a FAISS/Qwen-based RAG architecture for high-accuracy entity extraction and contextual Q&A.
- Owned the full deployment lifecycle, containerizing the system with Docker and exposing services via FastAPI, ensuring scalability and maintainability.
Shenzhen Chuangweida Electronics - Software Engineer Aug 2022 – Aug 2023
- Engineered backend microservices using SpringCloud for a high-traffic IoT platform, enhancing system modularity and resilience.
- Optimized real-time data synchronization between servers and thousands of terminal devices by implementing a robust messaging system with Netty, ActiveMQ, and WebSocket.
Beijing ZDZG Data Services - Data Analyst May 2021 – Sep 2021
- Automated a data pipeline for e-commerce analytics and deployed a TextCNN classification model on a Spark cluster, improving data processing efficiency.
Projects
LLM-Powered Conversational AI
- Developed a high-fidelity conversational agent by fine-tuning Mistral-7B using QLora on a film script.
- Implemented a RAG architecture with LangChain and ChromaDB, enabling the AI to leverage the entire script for deep contextual understanding and response generation.
Technical Skills
- AI & ML: PyTorch, LangChain, Transformers, Hugging Face, QLora, RAG, SpaCy, PaddleOCR, FAISS
- Languages: Python, SQL, Kotlin, Java
- Infrastructure: Docker, Spark, FastAPI, SpringCloud, ActiveMQ, MySQL, Git
Education
- M.S. in Computer Science, Yunnan University of Finance and Economics 2024
- GPA: 3.58/4.0 | Honors: Gold Medal, “Internet+” Innovation Contest
- B.Eng. in Internet of Things Engineering, Chengdu University of Info. Technology 2019
https://gawyn.de/2025/08/05/resume.en/