Kelvin Wu

• Lead and mentor 2+ engineers; manage technical roadmaps, sprint planning, and establish high-standard code review processes • Architected "Gold SKU API" (99.9%+ uptime SLO) centralizing product specifications across departments, reducing manual operational effort by 50% • Led comprehensive migration of 50,000+ legacy files from SVN to Git with zero downtime; doubled deployment velocity and reduced product release cycle by 35% • Designed GitLab CI/CD pipelines as Infrastructure as Code, eliminating release bottlenecks across engineering organization • Built internal LLM automation suite for end-to-end SDLC (ticket triage, code generation, testing, MR review), saving team 15+ hours per week • Architected AI conversational chatbot integrated with company documentation and internal APIs for real-time developer support and accelerated onboarding • Optimized SQL query patterns by 40%, significantly reducing server CPU utilization under peak load and improving search responsiveness

• Architected URL-shortening service achieving 3,500+ sustained requests per second with custom Key Generation Service (KGS) for distributed operation • Developed scalable distributed message queue system in 5 weeks, enhancing system efficiency and enabling asynchronous processing across microservices • Designed and published modular RBAC npm package adopted across multiple internal products, reducing access-related security incidents by 35% • Analyzed and optimized legacy SQL query patterns for enterprise applications, achieving 40% improvement in query performance and reducing CPU utilization • Collaborated with front-end engineers on live streaming e-commerce platform using Socket.IO and WebRTC for real-time bidirectional communication and peer-to-peer media streaming • Researched and implemented Redis Sentinel architecture for high availability with automatic failover and master-replica replication • Built low-latency TCP-based communication layer for inter-service communication optimized for high throughput and fault tolerance in distributed environments • Enhanced internal coding assessment platforms with AST (Abstract Syntax Tree) analysis to detect and prevent infinite loops during real-time code execution, maintaining 99.9% uptime SLO%, significantly reducing server CPU utilization under peak load and improving search responsiveness

• Developed a scalable distributed message queue system within 5 weeks, enhancing system efficiency. • Researched and showcased the Redis Sentinel distributed system, improving reliability. • Designed a high-performance URL shortening service, managing over 3.5k requests per second. • Collaborated with front-end engineers to implement a live streaming e-commerce platform using Socket.IO and WebRTC.

• Developed deep learning-based thermal compensation model for industrial machine tools, achieving 60% reduction in thermal measurement error through neural network analysis • Built automated data acquisition pipelines using Python-based neural networks to predict structural errors and enable real-time monitoring and compensation • Implemented AES encryption for thermal compensation core, securing sensitive measurement data and proprietary algorithms • Designed microcontroller firmware integrating memory, sensor, and display modules, enabling automated measurement workflows across manufacturing equipment