ibrahim mohamed

autonomous team vice lead/computer vision lead

• Job Details:designed autonomous navigation software using the python dronekit library and ardupilot firmware on cube flight controller. simulated the drone using gazebo with ros and ardupilot sitl for testing and development. developed a neural network using google’s inceptionnet architecture for alphanumeric symbol classification, achieving 99% training accuracy and 96% testing accuracy. engineered an effective pipeline for color classification of objects by integrating unet architecture to segment images and applying kmeans clustering, achieving an accuracy of 92%. solved object detection and localization challenge by integrating color classification, object detection, and alphanumeric classification to accurately pinpoint objects on the ground and calculate their absolute position. implemented a path planning algorithm using a* for obstacle avoidance and waypoint navigation.

ai intern

• Job Details:solved music classification problem by developing alghami, a music genre classifier using cnns and transfer learning in pytorch with 90% accuracy. gained experience with pytorch, tensorflow, gans, cnns, reinforcement learning, and optimization techniques. researched audio signal preprocessing and feature extraction using librosa and torchaudio.