Since I worked on "new tools" based on Computer Vision and ML, I had the opportunity to move teams from the Quality Assurance to the company’s innovation laboratory.I worked on the benchmark tools and test automation, as well as other projects.Most of the projects concern "Video Enhancement". One of these was Video Super-Resolution. Since Shadow centers on video streaming, the idea is to reduce the required bandwidth by sending a small image and upscaling it client side. This project is on-going, and we have achieved satisfactory results, both in terms of quality and performance.My other projects include cursor movement prediction, user habit pattern prediction, and a side project involving social media analysis that feeds into an other project on user satisfaction and churn prediction.

When I joined the company as a Quality Assurance Engineer, there was no Quality Assurance Team.I helped to define the processes and test cases.Since Shadow offers a "full computer" experience, many functionalities needed testing, most of which where not complex but repetitive and time consuming. I opted to use python scripts to automatise those tests.In order to avoid environment corruption, all the tests ran on local machines and interacted with the virtual machines through an application. This method involved one major challenge, as we relied on the live streamed image to navigate and interact with the virtual machine. I designed a solution leveraging computer vision, firstly by using OpenCV and then adding some ML algorithms.

Creation of a database of reliability studies of electronic functions. The mission is to retrieve data, check if they are suitable (if all the information has been entered correctly) create the database of these function blocks (Use of MS Access) as well as demonstrate a probability of correct operation (for the product and from a security point of view). When we do a Fides Life Profile some components may be used several times. In order to gain some time, we want to import the "classic" blocs directly in the Fides Excel Sheet.