Lead researcher on the development of a multi-sensory data fusion platform using RF-ID, Lidar, low-visible light vision camera, and deep learning to optimize faster RCNN target detection algorithms on SOSCIP GPU-platform for night vision and location estimation of targets and assets in dark underground environments.Developed RRSI-based RF-ID location estimation system and RF-ID tag based on ESP32-Wrover (Hardware design using Altium) with indoor localization using multiplatform technologies such as WIFI, BLE, XBEE, and UWB.Built machine learning models in Caffe and TensorFlow using proposed feature extraction with implementation on GPU.Designed, implemented, trained, and optimized novel algorithms that run at low-latency production systems. Designed RF amplifiers for underground communication at schematic and PCB level

Developed a computer vision-based indoor navigation system for autonomous UAV surveying of fuel channels in power plants using ROS and OpenCV in python.Developed a portable light detector-based gamma counter. Reverse Engineered the control panel equipment for simulator compatibility.Improved and delivered reliable firmware for various signal processing applications such as test jigs through continuous integration, automated testing, and code reviews.Developed complex, high performance FPGA firmware and software for data movement, I/O interfacing, and embedded signal processing.Developed a vision-based indoor navigation algorithm for autonomous UAV surveying of Fuel Calandria in Candu plants.Designed the I/O subsystem for Control Room Simulator of Wolsong power plant in South Korea.Designed & developed the voltage comparator and trip conditioner module for Embalse power plant in Argentina.

Developed image processing and computer vision algorithms for image recognition, image classification, object detection, object tracking, image/video semantic understanding, image/video detection, and image/video segmentation.Development of a novel image and video quality assessment method for de-noising signals, images, and videos.Development of a Support Vector Regression (SVR) based technique for quantifying imagemodifications using information-theoretic distances as features.Built tools to train deep learning models for image and point cloud data and test datasets for different image and video quality assessment algorithms for biomedical applications.Developed an HDR tone-mapping parameter-tuning algorithm for visualizing medical images on conventional displays for mammogram imaging.