Artificial Intelligence Intern — Centre for Development of Telematics (C-DOT)
Developed an AI-based real-time video analytics system focused on multi-object detection and object tracking. Worked on computer vision pipelines that include metadata processing and stream optimization for intelligent surveillance outputs. The work involved preparing/handling video frame outputs and associated tracking metadata for downstream analytics. • Implemented multi-object detection and tracking logic using NVIDIA DeepStream and GStreamer. • Processed and optimized metadata associated with tracked objects for surveillance use cases. • Improved stream handling and analytics efficiency for real-time performance. • Built components of an end-to-end video analytics workflow for object tracking.