Muhammad Azan

I specialize in computer vision and machine learning, delivering real-time AI solutions for object detection, pose estimation, action recognition, and anomaly detection. I build production-ready models using YOLO, TensorFlow, and PyTorch optimized for accuracy and speed.

As a freelance CV and AI specialist, I've delivered end-to-end systems—from dataset annotation and model training to deployment with Docker, SageMaker, GCP, and Flask/Streamlit interfaces. My projects include sports analytics, PPE and workplace safety systems, and fast video-processing pipelines, often achieving >90% detection accuracy with low-latency inference. I also consult on model tuning, API integration, and scalable cloud deployments.

I enjoy turning research into practical products and sharing insights through technical content and demos. I'm motivated to join teams that value applied AI, production deployments, and measurable performance impact.