Deploy highly precise visual intelligence models designed to operate directly on constrained edge hardware in real-time. I specialize in designing and optimizing computer vision pipelines using state-of-the-art frameworks like YOLO, EfficientNet, and Vision Transformers. My core capability is multi-modal sensor fusion—integrating raw visual streams with gas sensors, near-infrared data, or metadata via early fusion to maximize accuracy while achieving sub-150ms inference latencies on portable edge processors.

Custom dataset preprocessing, augmentation, and model training

Advanced model architecture configuration (YOLO, CNNs, Transformers)

Multi-modal sensor fusion system integration

Model quantization, optimization, and pruning for edge compilation

Real-world field deployment testing and performance profiling