Designed and optimized real-time mapping and localization algorithms using Mid-360 LiDAR, multi-sensor data fusion, and fast_lio for autonomous systems. Developed robust Python/C++ pipelines for odometry preprocessing and secure UART-based communication. Integrated MAC-based hardware verification to ensure software integrity and deployment security. Collaborated with researchers to fine-tune algorithms, enhance system performance, and enable accurate navigation in dynamic environments.

• The project utilizes Jittor in Python and convolution kernel to apply computer vision deep learning techniques based on texture extraction to assist in the diagnosis of skin diseases.• Enhanced training dataset quality and optimized performance through preprocessing. • Configured a robust deep learning environment and executed efficient data training processes. • Conducted algorithm evaluations for efficiency and accuracy, contributing to optimal solutions. • Collaborated on applying trained models in practical cases, receiving accurate feedback, and addressing medical image dataset limitations.

• The project used deep learning algorithm for cell segmentation in multiplex tissue image. • Established a high-throughput imaging analysis pipeline for analyzing multiplex tissue images. • Created a pre-processing algorithm to remove low quality images and enhance image quality. • Implemented a deep learning-based cell instance segmentation algorithm and optimized it.