FAST is an ultrasound procedure that aims to evaluate the presence of free fluid in heart and abdominal areas. This research investigates the possibility of using deep neural networks to automatically assess trainees’ FAST skills. A deep convolutional neural network was trained to assess FAST skill level from time-series motion data, represented as sequences of transformation matrices, obtained from scanning four regions of a human body. The data was gathered from 32 users with different proficiency levels: novice, intermediate and expert. K-fold cross-validation protocol was used to evaluate the network’s performance. The results show that the proposed deep convolutional neural network can classify proficiency with high accuracy. The research findings demonstrate the potential of using deep learning-based skill assessment tools in medical training.