At Qualcomm India Pvt. Ltd., I developed object-visualization smoothing for automotive software using rule-based methods, Kalman filtering, and LSTM temporal models. I also implemented object clamping, class hysteresis, and yaw-based smoothing.
I automated crash-dump analysis with Python, extracting register values, flags, and diagnostic data for graphics failure investigation. I also developed graphics routines in C++ and DX11 and investigated graphics hazard cases.
For my Network Intrusion Detection System project, I trained and evaluated machine learning and deep learning models to classify benign traffic and attack categories. The project achieved 98.78% accuracy with Random Forest and 97.18% with LSTM, and used explainable AI and feature selection to identify important features.
I designed a hybrid intrusion detection framework to identify previously unseen attacks, and implemented an encrypted chat application using AES-256 and RSA. As captain of Team Rotor FPV, I led development of autonomous, racing, and freestyle UAVs, including work on path planning and object detection.

