Ahmad El-Sherif

• Was responsible for generating a synthetic dataset using the 3D modelling tool ’Blender’ for in-cabin car scenarios and benchmark testing it for different tasks including pose-estimation and object detection for RGB and depth images. • Worked on a variety of tasks from segmentation masks and bounding boxes annotation, and creating and managing the department and dataset websites using PHP and Wordpress.

• Worked in the research and development of tools and platforms in the Comfort Driving Vision (CDV) department that included working on a Point Cloud visualizer and data extractor from different sensors. • Used different technologies to build and develop AI-based tools to automate and improve the testing and validation cycle such as an object classifier and a part-segmenter for vehicles.

• Implemented a Deep Neural Network using pytorch that is able to recognize and segment different road objects given 3D Point Clouds generated from LiDARs and Radar sensors in addition to using simulator-generated data for training. • Using multi-layered height-maps with dilated convolutional networks for the segmenter.

Implemented a peer-to-peer social networking service for students to find talented tutors with a user-based ranking system. The project was deployed on the then newly launched IBM cloud server Bluemix.

Collaborative Intelligence; Multimedia Data Mining, Very Deep Learning; Applications of Artificial Intelligence; Case-Based Reasoning; Embedded Intelligence.

Machine Learning; Algorithms; Embedded Systems; Structure of Programming Languages; Operating Systems; Databases; Software Engineering