Some roads are drawn on maps. Others are drawn in will.
My Tech Stack
Projects
Used Car Picker Web Application
Oct. 2024 – Dec. 2024
- Developed full‑stack web app with Node.js, MongoDB & D3.js to visualize 5,000+ vehicle records.
- Built RESTful API with optimized MongoDB queries & Redis caching, reducing response time from 2–3s to <200ms.
- Designed responsive UI with interactive filters and price‑tracking algorithms for vehicles 15% below market value.
- Implemented error handling & data validation, cutting crashes and improving UX with clear feedback.
GFW Research (Master’s Thesis)
Sep. 2023 – Present
- Analyzed China’s Internet censorship using custom Python tools & network analysis across millions of users.
- Built Docker‑based distributed probing infrastructure for 10,000+ domains, collecting 500 GB+ of traffic.
- Quantified censorship trends via DNS errors, IP patterns & packet inspection to map filtering nodes.
- Explored socio‑political impacts of “information cocoons”, contributing to academic discourse on free flow of knowledge.
macOS Music Player
Jan. 2023 – May. 2023
- Built native SwiftUI app with Core Audio for lossless playback and custom UI components.
- Trained CoreML model for personalized playlists (65% → 85% relevance) using collaborative filtering.
- Achieved 90% test coverage with XCTest and GitHub Actions CI/CD, reducing integration issues by 70%.
- Optimized audio pipeline to cut CPU usage by 40%, extending battery life on older Macs.
- Published DocC docs & user guides; community contributed 5+ plugins without altering core code.
ML‑Powered Email Security Extension
Aug. 2022 – Dec. 2022
- Built Chrome extension with JS frontend & Python (scikit‑learn) backend for phishing detection.
- Extracted 15 key features; ML pipeline achieved 92% accuracy vs. 75% default filters.
- Reduced inference time from 800 ms to 150 ms per email, enabling real‑time protection.
- Secured API communication with auth & encryption, protecting user content.
- User testing with 50 participants; UI improvements cut false positives by 60%.
WeChat Admissions MiniProgram
Aug. 2021 – May. 2023
- Built WeChat MiniProgram (WXSS/WXML + Node.js & MongoDB) for international student recruitment.
- Implemented secure auth & RESTful backend; grew to 1,000+ DAU despite time‑zone differences.
- Applied agile sprints; delivered 35+ feature requests based on stakeholder feedback.
- Developed analytics dashboard for admissions staff, increasing conversion rates by 24%.
- Trained staff & handed over detailed docs, ensuring sustainable platform maintenance.
Home Datacenter Project
Jun 2023 - Aug 2023
- Deployed Proxmox VE virtualization on underutilized hardware, consolidating 5 critical VMs (3 Windows, 2 Linux) and saving ¥50,000 in hardware costs by eliminating physical servers.
- Built Docker Compose-based dev environment on Ubuntu with Ollama AI, Nginx, GitLab, NAS, and proxies. Automated setup reduced deployment time from 3 days to 30 minutes.
- Implemented CI/CD with Jenkins, GitLab CI, and GitHub Actions, synchronizing code between internal and external repos, reducing integration delays and merge conflicts by 75%.
- Configured network security with firewall rules and proxies for secure file sharing and remote access, achieving successful ISO 27001 audit.
- Standardized dev environment with VM templates and automated bootstrapping, reducing developer onboarding time from 5 days to 2 days (60% faster).
Hengyi Li
DevOps Engineer and Software Developer with expertise in CI/CD pipeline optimization, containerization (Docker, Kubernetes), and system administration. Skilled in Python, Bash, and JavaScript development with experience in automating infrastructure and implementing high-availability solutions. Proven track record of reducing deployment times by 80% and ensuring 99.9% service uptime through robust system architecture.