Summary
Xiaoyi Li is a Senior Software Development Engineer with 12 years at Tencent, specializing in cross-platform systems, multimedia, and AI inference for audio/video applications. He architected and shipped a cross-terminal file transfer SDK for QQ that achieved >99.9% single-file transfer success and developed live-streaming SDKs and monitoring pipelines. More recently he led development of XNN, a portable AI inference engine optimized per platform (NEON/AVX/Metal/OpenCL/D3D/CUDA) powering features like gesture recognition, portrait segmentation, denoising, and real-time beauty for flagship Tencent products. Comfortable across C++, Python and Objective-C, he blends low-level performance optimization with production-grade observability and quality reporting. Based in Shenzhen, he pairs a Sichuan University computer science background with deep experience making high-throughput media pipelines reliable at scale. An understated strength is his focus on platform-specific micro-optimizations that squeeze real-world latency and accuracy gains across desktop and mobile deployments.
12 years of coding experience
Bachelor of Science (BS), 计算机科学与技术, Bachelor of Science (BS), 计算机科学与技术 at 四川大学
Chinese, Chinese