Xin Jiang

AI Application Engineer at Lenovo CTO Org (Foundation Model Team), specializing in scalable model inference systems and cross-platform AI deployment. Designed and built a high-performance C++ inference framework supporting multi-model scheduling, lifecycle management, and heterogeneous backends (CUDA, ROCm, Intel, ARM) across Windows and Linux. Led on-device deployment of 7B LLMs and Stable Diffusion 3, optimizing latency, concurrency, and system efficiency. Drove end-to-end inference performance optimization (ONNX), achieving ~89% improvement over ONNX Runtime via backend integration (DirectML, OpenVINO, NCNN, MNN), quantization, and preprocessing. Developed lightweight Android inference solutions and enabled full-stack pipelines integrating Kotlin (Quantum Core) with C++ services. Contributed to multiple AI products including Lenovo AI Now, Xiaotian, Tianxi, and Qira. Recognized with FY24/25 VP Semi-Annual Team Excellence Award. Inventor on patents in cross-language inference frameworks for heterogeneous computing and advanced RAG strategies improving multi-chunk reasoning and response quality.

Full-Stack Software Engineer at Lenovo Research (Edge AI Team), focused on scalable edge AI systems and industrial AI deployment. Built a few-shot model training and inference platform with a Python async backend and Vue.js frontend, designing a multi-protocol communication framework (HTTP / WebSocket / Socket / MES) capable of handling 4 concurrent streams and processing 2GB image data within 5 seconds. Optimized inference pipelines with parallelization and GPU acceleration, reducing industrial defect inspection cycle time from >10s to ≤4s. Led integration between training and inference systems and supported end-to-end deployment, including system integration, on-site delivery, and POC validation for enterprise clients such as CATL, ATL, and Amphenol, enabling stable production rollout. Developed a cloud–edge–end encryption and authorization system, including a C++ crypto library and a Golang-based licensing platform. Enabled secure, cross-platform deployment across C/C++, Python, and Golang environments on Windows and Linux. Recognized with FY23/24 VP Semi-Annual Team Excellence Award and VP Project Instant Awards.

Contributed to the development and maintenance of the user knowledge base and after-sales platform, supporting feature implementation and software testing. Developed and maintained internal tools for development and testing, and translated, organized, and managed internal English technical documentation to improve team efficiency and collaboration.

Built and ran basic predictive models using TensorFlow to analyze biomedical data. Cleaned, visualized, and conducted basic statistical analysis on research datasets using Python and R. Explored AI applications in biomedical research, including classification tasks and biomedical text extraction, to deepen understanding of AI’s role in biomedical research workflows. Collaborated with U.S.-based managers and cross-country team members through regular video meetings and technical documentation exchange, strengthening intercultural communication and teamwork skills.