● Developing real-time multi-object tracking (MOT) algorithms in C++, leveraging FMCW Doppler velocity to improve tracker performance in challenging environments.

● Optimizing embedded perception pipelines for custom SoCs using NEON intrinsics, fixed-point math, and memory-efficient data structures to meet strict compute budgets.

● Implementing and tuning classical computer vision algorithms (ground segmentation, clustering) for 4D LiDAR point clouds.

● Building Python analysis tools to visualize internal algorithm states, enabling faster debugging and validation.

● Developed a realtime Camera-LiDAR fusion perception stack for traffic monitoring applications using C++, Python, and ROS 2, resulting in an 8% increase in object detection and tracking accuracy over single sensor solutions.

● Created a custom unsupervised 3D pointcloud detection and tracking module combined with computer vision for tracking vehicles and pedestrians.

● Conducted end-to-end system testing and verification of functionality with real-time live data directly from sensor hardware.

● Developed production C code for new features and bug fixes. Performed system-level and vehicle-level testing and validation.

● Created comprehensive automated unit test suites with C++ for new and existing software components, ensuring the reliability and quality of the codebase with

90% test coverage. Increased automation of regression testing by 20%.

● Designed and assembled a remote testing platform for ECUs with CAN, LIN, and Ethernet using CANalyzer, reducing testing time by 30%.

● Designed and 3D Printed a mobile robot platform using SolidWorks and Cura. Selected motors, drivers, and electrical components.

● Developed ROS nodes for wireless robot control and initiating RADAR data collection.

● Diagnosed and solved hardware and software issues with Wireshark and an oscilloscope.