Developing Generative AI tools to protect some of the most popular brands from phishing and brand impersonation attacks.

Developed data streaming pipelines for psychophysics datasets to estimate behavioral cues of pedestrians in autnonmous driving scenarios.

Aggregation and analysis of spatio-temporal features to encode human behaviors suitable to train a deep learning pipeline

Designed multi-task neural network architecture to predict pedestrian intent aggregating features from psychopysics data pipelines.

Automating a Heavy Commercial Truck and a Subcompact Crossover SUV by developing Operational Design Domain specific Object and Event Detection and Response solutions under the Perception Subsystem by leveraging the power of Deep Learning.

• Worked on integrating Deep Learning with Computer Vision.

• Learned how to train Convolutional Neural Networks using Keras with TensorFlow backend.

• Worked with different state of the art CNN architectures like Inception, VGG and SSD.

• Worked with TensorFlow Object Detection API for object detection tasks.

• Spearheaded the interfacing of client project SnapCam on EDGE IOT devices like Raspberry Pi &

NVIDIA Jetson Nano for optimized inferential object detection.

• Execution of cross-platforming for Face Recognition Data Base module of product DeepSight from Linux to Windows by configuring SQLITE3 library, along with existing FAISS library.

Worked on different Object Tracking techniques, Feature matching, Object Detection using Cascaded Classifiers, Occlusion Handling and motion analysis under the guidance of Dr. Sanjay Mathur, Professor, ECE department, G.B.P.U.A.&T.