At Zeus Artificial Intelligence, I architect and optimize real-time computer vision pipelines for an enterprise IoT cloud platform serving agricultural and industrial monitoring. I process multi-camera video streams and integrate vision outputs with cloud backends and IoT networks.
I fine-tune object detection models across the YOLO ecosystem and optimize inference for NVIDIA Jetson and AMD Ryzen AI hardware using TensorRT, ONNX Runtime, and quantization.
At L2M Rail, I developed and deployed anomaly detection and object detection models for railway infrastructure inspection. The models boosted prediction accuracy by 12%, and I also curated and preprocessed a dataset of more than 100,000 images.
Earlier at L2M Rail, I worked as a Machine Learning & Deep Learning Intern, optimizing edge ML models and evaluating predictive agent behavior. My projects include an automated handwritten answer script evaluation system using OCR and NLP.
