At Inferigence Quotient Pvt. Ltd, I built computer vision data and annotation pipelines, including COCO-to-DOTA conversion and image tiling for oriented bounding box detection. I also integrated GradCAM analytics into mmYOLO workflows to support model interpretability.
I developed a benchmarking toolkit for NVIDIA Jetson that analyzed hardware telemetry and compared TensorRT model performance across FP16, FP32, and INT8 modes. I also worked on geospatial traffic simulation and trajectory generation for Electronic Warfare scenarios.
At PESU Center for IoT, I led development of the Pest Detection Rover, deploying a CNN-based pest detection system on Raspberry Pi and ESP32. I co-authored research presented at the IEEE Conference on IoT, and my journal publication is under peer review.
