Worked on research projects towards Machine Learning and SLAM for L3+ level of Autonomous Driving solutions Highlights: • 4D Labeling Designed and implemented a semi-auto labeling solution combining Structure from Motion and Machine Learning (Neural Radiance Fields) to build 3D models for multi-sensor driving data, delivered results for over 350km data in five cities, resulted in reducing the labeling time by 50%, and continuously integrating and delivering new data • LiDAR SLAM Developed LiDAR algorithm with existed Visual SLAM framework, working in ROS environment offline and embedded system in real-time for various vehicle models including demo cars and production vehicles • Evaluation System Identified standards and developed tools for evaluating the performance of SLAM modules and object detection modules, used to test performance among internal projects • Driving Data Simulation Implemented fisheye camera simulation in Unreal Engine 4 based driving simulator (CARLA), generated multi-sensor datasets in 10 city maps with seven weather including various objects such as vehicle, pedestrian, bicyle. Generated data served the training of Machine Learning models • Patent Applied five patents about Machine Learning and SLAM, three of them has been accepted and published, two under processing

South Australia, Australia Instructed private 1-on-1 tutorials for object-oriented programming in Python - 2 students scored 70+/100 in their programming course

South Australia, Australia • Optimized the moving kinematic of omni-wheel unmanned ground vehicles (UGVs) by redesigning the control system • Developed controlling tools with GUI for robot movement and robot arm control