- Experienced with State of the art Object Detection Deep CNNs. ranging from ResNets, to EfficientDet.
- Hands-on with python, pytorch, and SOTA research platforms(e.g., detectron2 …)
- Implemented a family of modern, scalable and efficient damage detectors with AP50 53.6, F1-score 56.5 % and Inference Time 200 images/sec.
- familiar with different mechanisms to enhance the process of training and deploying a network (e.g. Auto-Augment, mixed precision training, ensemble methods, TTA, …)
- Contributed to several Open-source GitHub repositories e.g. google/autoML, bbaug, … .
Related links :
Sadra Naddaf
Sr. Machine Learning Engineer
My research interests include Machine Learning, NLP, and Deep Learning.
Publications
An Efficient and Scalable Deep Learning Approach for Road Damage Detection
Pavement condition evaluation is essential to time the preventative or rehabilitative actions and control distress propagation. Failing …
