Prince Choudhary

I’m an AI/ML Engineer with 3+ years of experience developing, training, and deploying production-grade artificial intelligence solutions. I focus on building efficient, reliable AI that solves real-world problems and drives business impact.

In my current role, I built a Real-Time Facial Recognition & Person Tracking system—training a YOLO model for face detection, integrating MediaPipe for face quality scoring, and filtering frames to improve throughput. I optimized the pipeline by encoding only filtered unknown faces for AWS Rekognition, reducing API calls by 40% as cost, and I quantized models to ONNX for Jetson Orin Nano while improving real-time performance using multithreading and RTSP buffering.

I also develop end-to-end vision and streaming systems, including an RTSP video streaming pipeline using FFmpeg and MediaMTX for live monitoring and video analytics. Previously, I implemented YOLO-based QR/barcode detection (improving accuracy by 16%), built GeoSyncVision panoramic mapping by aligning satellite coordinates to video frames, and created an automated trespass detection system with YOLOv8, line-crossing logic, and polygonal ROIs—reducing system load by 20% with adaptive frame skipping. I further build secure AI apps like a RAG chatbot with semantic search and LLM guardrails (prompt injection protection and hallucination detection).