Deep Radar Object Detection for Autonomous Driving

Deep Learning-based Compression and Quality Assessment of Point Clouds- Lossy compression of point cloud geometry using Convolutional Neural Networks (CNNs)- Folding-based lossy compression of point cloud attributes- Lossless compression of point cloud geometry using masked CNNs (co-supervised intern)- Deep learning-based point cloud quality assessment (collaboration)- Deep learning-based LiDAR point cloud geometry compression (internship)

Deep Learning-based LiDAR Point Cloud Compression

- Thumbnail Selection From Movie Trailers for the Microsoft StoreDesign, implementation and production deployment of a scalable thumbnail selection system for the Microsoft Store's movies using CNNs.- Books Rating Classification for the Microsoft StoreProof of concept book classifier for the Microsoft Store to detect rating misclassifications using LSTMs.

Ruby on rails application development, system administration, mail server administration

- Moonshot Analytics, Analytics and insights from raw sales data- Allianz, « Visio Zen », Remote medical assistance with WebRTC video calling- Groupama, Status tracking of remote devices by aggregating and analyzing MQTT data- Zodiac Aerospace, Generation of wiper models based on available parts and requirements (size,architecture...) : Built a custom combinatorial search algorithm- Axema, Farm trailer stability prediction in different situations (wind, slope, tyre pressure...)

Lead architect and developerNOE (Groupama), a social network and remote assistance system for the elderlyWebRTC Video calls, notifications, cross-platform, cross-browser,responsive design, continuous integration