Computer Vision Engineer

I worked as Computer Vision Engineer for 3 years at Heuritech, first as an intern and then full-time.

My main task as a computer vision engineer was to implement and improve existing deep learning models with TensorFlow. Those models performed:

• object detection (using Faster-RCNN)
• color detection (using CNN and attention mechanisms)
• multiclass and multi-label classification (using CNNs)

I also implemented evaluation procedures of our computer vision pipeline to generate actionable insights for the business team.

As a software engineer, I maintained the team’s repository by:

• setting up and maintaining a CI that performed: docker image building, unit testing with coverage computation, linting, documentation generation and upload
• implementing documentation generation using Sphinx
• managing the life cycle of our docker images
• managing the mono repository internal dependencies using Poetry

While working at Heuritech, I worked on 2 open source projects:

• Garbage detection on videos with Surfrider - Adapted object detection technologies to perform garbage detection and tracking on videos of river banks
• GitLab Slack Notifier - A tool first developed for internal use that we open sourced later