基于 3D视觉感知、大模型Agent 与 数据驱动技术，为全球企业构建灵动高效的智能化体系。 Building agile and efficient intelligent systems through the integration of 3D Vision, AI Agents, and Data-Driven Technology.
基于 LangGraph + MCP 架构构建的下一代智能语音问答系统。专为复杂业务场景设计，支持多语种无缝切换与情感识别。 Next-gen intelligent voice Q&A system built on LangGraph + MCP architecture. Designed for complex business scenarios with seamless multi-lingual switching.
拥有高智商大脑与动听声音的交互式数字人。整合 LLM 与 SadTalker 技术，提供低成本、高逼真的数字员工解决方案。 Interactive digital humans with high IQ. Integrating LLM and SadTalker technology to provide cost-effective, hyper-realistic digital workforce solutions.
基于改进 YOLOv8 的牙片图像分割系统。引入空间注意力机制，解决 X 光片对比度差的难题，精准辅助医生诊断。 Dental X-ray segmentation system based on improved YOLOv8. Introduces spatial attention mechanisms to solve low contrast issues in X-rays.
覆盖从“猜你喜欢”到“精准化推荐”的全链路智能引擎。广泛应用于旅游规划、商业竞价与供应链筛选。 A full-link intelligent engine covering everything from "Guess You Like" to precision targeting. Widely used in travel planning and commercial bidding.
自研基于 Triton 的多 GPU 推理与运维平台。不仅是运维工具，更是大模型微调与异构算力调度的效能引擎。 Proprietary multi-GPU inference and operations platform based on Triton. An efficiency engine for LLM fine-tuning and heterogeneous compute scheduling.
Magitech 是一家以人工智能驱动的软件服务公司，致力于提供高效、精准、可落地的 AI 解决方案。我们深耕各行业 AI 应用，快速构建 AI 系统，帮助客户实现业务目标，提升效率和竞争力。 Magitech is a technology company dedicated to providing high-quality software services driven by artificial intelligence. We deliver efficient, precise, and practical AI solutions, exploring applications across various industries. By rapidly building AI-based systems, we help clients achieve their business goals, improve efficiency, and enhance competitiveness.
前 Nokia/Salesforce 高管，NUS 工商管理毕业。 Former Nokia/Salesforce Exec, NUS MBA graduate.
名校硕士，前京东/Nokia 研发负责人。 AI Master, Former R&D Head at JD.com/Nokia.
NUS 机器人视觉交互博士，IEEE Fellow 审稿人。 NUS PhD in Robot Vision & Interaction, IEEE Fellow Reviewer.
