At Samsung Electro Mechanics, I designed and deployed production computer vision systems for industrial surveillance, achieving 95%+ accuracy in fire, smoke, and PPE detection.
I worked across the AI lifecycle, from data strategy and model training to optimization and factory integration. I also developed C++ inference pipelines and packaged models as DLLs for low-latency execution on constrained devices.
For real-time performance, I implemented TensorRT-optimized models and contributed to model benchmarking and iterative tuning. My work focused on meeting operational requirements and maintaining reliable production systems.
Earlier, at AsterQuanta, I delivered a real-time facial recognition platform and contributed to its cloud-native AI pipeline using AWS services. My project work has also included worker fall detection, in-vehicle surveillance, and crop quality prediction.

