Tech lead on large language model (LLM, Multimodal, Reasoning model) research, SW-HW co-design, model optimization for training / inference on GPU and Custom AI Silicon, and generative AI system design & innovations.

Machine learning framework development, large language model (LLM) acceleration on CPU/GPU, ML algorithm design, CNN/RNN/LLM model optimization, GPU kernel development, model serving/deployment for computer vision, speech, NLP applications in cloud/edge that serve billions of users everyday.

Led the effort of wireless system design (Bluetooth/WiFi/LTE), architecture definition, baseband algorithm development, software implementation, system integration and performance tuning on SoC/DSP/FPGA. Responsible for algorithm modeling in C++/SystemC, link-level simulation, High Level Synthesis (HLS) to RTL, software/hardware integration, co-design/co-validation, cross-layer optimization and debugging. Worked closely with other core engineering teams on mixed signal simulation, link budge analysis and end-to-end performance optimization. Fixed critical issues and ensured on-time delivery of massive products to customers.

Led firmware and software teams on advanced algorithm design and software development for mobile WiMAX and VDSL products. Optimize OFDM-MIMO algorithms and MAC-PHY cross layer procedures. Implemented L1/L2 protocol stack software on embedded multi-core DSP SoC. Simulated end-to-end system performance in C++/Matlab.

Led DSP team on baseband development for 802.16e mobile WiMAX products. Designed advanced signal processing algorithm and software on DSP/FPGA platform. Led the effort of SW/HW integration, debugging and field testing. Helped the company to deliver the world's first SDR based WiMAX CPE and base station products.