Houssem’s research involves building smart video search engines that use new ways to search. For example. in a project on self-driving cars, he is designing a search engine for traffic scene videos that can conduct searches based on text (e.g. "car changing lanes") and even sketches depicting icons of objects  and arrows. He also works on new technologies for education. At the moment, he is working on enabling more intuitive access to the popular Khan Academy platform by allowing users to input a sketch and conduct a search based on that. Here, a user would draw a sketch on a tablet (e.g. a figure, math or chemical equations, diagrams), and the search engine retrieves the relevant Khan Academy videos. In all his projects, the main challenge is to draw high level semantics from videos using the right AI tools.