Xuechun Jin

CS MS @ UW–Madison (PMP) | Seeking 2026 SDE Internship

Madison, Wisconsin, United States

About

MS CS candidate at UW–Madison with a strong foundation in software engineering and AI/data systems. Skilled in backend development, databases, distributed systems, and applied AI/ML (IR, NLP, LLMs, Computer Vision). Experienced in building scalable platforms with measurable performance gains and leveraging AI-assisted tools to accelerate development. Seeking Summer 2026 internship opportunities in software engineering, data systems, or AI/ML.

Experience

  • Software Engineering Intern at LYNXdigital
    May 2025 - Aug 2025 · 4 mos

    Built a real-time monitoring platform, cutting BI costs by 30% and eliminating reporting delays. Migrated multi-source data into PostgreSQL with indexing, ~300× faster queries. Developed Node.js/Express APIs with Redis caching, sustaining 500+ concurrent requests <200ms. Designed React + Plotly dashboards for live workforce visibility. Adopted AI-assisted development (Copilot, ChatGPT), halving iteration time and improving code quality.

  • Independent Project (7 mos)
    • Web Information Retrieval Engine
      Jan 2025 - Mar 2025 · 3 mos

      • Developed a scalable search engine processing 56k+ crawled web pages. • Implemented inverted index with optimized merging, reducing memory usage and achieving <300ms query latency. • Integrated semantic embeddings + RAG (OpenAI API) for intelligent, LLM-powered retrieval. • Improved relevance by ~15% on 20+ benchmark queries.

    • 3D Stereo Reconstruction
      Sep 2024 - Dec 2024 · 4 mos

      • Designed a computer vision pipeline with structured-light coding and stereo correspondence for depth estimation. • Improved reconstruction accuracy through point cloud refinement and mesh smoothing, reducing noise and enhancing geometric fidelity. • Built full Python pipeline (OpenCV, NumPy, Matplotlib) to generate interactive 3D models, enabling scalable visualization and analysis.

  • Guide at Tiangong University
    Sep 2020 - Dec 2021 · 1 yr 4 mos

    • Collaborated in a team of 6 to design a traction-type hexapod guide robot for visually impaired users. • Implemented locomotion algorithms (tripod gait, stair-climbing) with optimizations ensuring efficient and stable performance on complex terrains. • Integrated YOLOv3 stereo vision with ROS, achieving >99% stair detection accuracy in real-world stair-climbing experiments. • Recognized as a National College Student Innovation & Entrepreneurship Project.