At Antei, I developed Vision Language Models for multimodal AI applications and optimized YOLOv11 for real-time action and object detection, improving detection performance by 45%. I also implemented DeepSORT-based multi-object tracking and reduced computational cost by 10% through data preprocessing and fine-tuning.
With DJS Antariksh, I developed data acquisition systems, control algorithms, and microcontroller integrations for autonomous rovers. I implemented robotic control using ROS2 and C++, and contributed to competition finishes including third place twice at the International Rover Challenge.
