/about

I'm a deep learning engineer specialized in computer vision based in Paris, FR. I enjoy experimenting from latest research papers and designing engaging and creative AI applications.

Contact me

Skills

pythonpytorchtensorflowfastapiflaskdjangotypescriptjavascripthtmlcsstailwindreactreduxvitesqlitepostgresredisrabbitmqmongodbdockerkubernetesgitlinuxcpprbashlatexvscodeawsgooglecloudfigmamarkdownunityunrealblenderphotoshopillustratorpremierexd

Work Experience

  1. 2021 — Present
    Deep Learning Engineer
    Airbus Defence and Space

    Lead the research, development, and architecture of deep learning products using object detection, identification, keypoint extraction, pose estimation and similarity search in large collections of satellite imagery. Implements latest research papers and techniques to train models on frugal datasets including but not limited to metric learning, self-supervised learning, transfer learning and active learning.

  2. Apr — Sep 2020
    Deep Learning Engineer
    Thales Defence Marine Systems

    Build and trained deep learning model from research publications for multi-modal weather nowcasting to help Unmanned Underwater Vehicle (UUV) navigation and route planning. Engineered a full-stack service for nowcasting to generate predictions using a Postgis database and more than 100+ data sources.

  3. May — Aug 2019
    Computer Vision
    IGN France

    Leveraged image matching algorithms and applied photogrammetry techniques to generate accurate point clouds from multi-spectral drone images. Conducted 3D classification to map vegetation in the French Alps in collaboration with local authorities.

  4. 2019 — Present
    Photographer & Designer
    Freelance

    Design illustrations for Noovae, a 3D virtual reality startup to help them communicate on world heritage. Collaborated as a photographer for Vikvize, a Norwegian counseling agency to capture impactful visuals to support their services.

View more on Linkedin.

Selected Projects

  1. geolabel-maker
    geolabel-maker

    Geolabel Maker is a tool used to generate labels from satellite or aerial images and vector geometries.

    • 20+
  2. nets
    nets

    Recreating PyTorch from scratch, using Numpy only. Implements FCN, CNN, RNN layers and supports CPU or CUDA computation.

    • 26+
View more projects.

Contact

If you have any questions or just want to say hi, feel free to reach out to me at any of my socials above. I'm always open to opportunities and new connections!