Machine Learning Intern
Allur
Jun 2026 — Sep 2028• Built real-time CV pipelines on high-load RTSP streams - detection, tracking, and video processing on production sites, not just benchmarks. • Optimized inference with TensorRT to reduce latency and increase throughput in production deployment. • Integrated MinIO to store events and video artifacts from real-time streams. • Built a worker efficiency monitoring system for an automotive assembly line: detection and tracking of workers. • Prepared and labeled datasets in CVAT and Roboflow, applied augmentation for model fine-tuning.