Computer Vision Developer in Paris, France

Created an ML pipeline that rectifies pictures of license plates and recognizes their numbers. I selected relevant existing models and improved their accuracy and their latency. I then developed a production environment in AWS.

Managed a team of developers to build a voting platform under a heavy workload. The platform comprises a React-based front end hosted on Netlify and several microservices used for the back-end using AWS. I carefully monitored errors. Activity monitoring has been essential for the project's success.

Helped my clients automate processes with their CRM. I developed new user interfaces based on Airtable and React. Then the interfaces were connected to the CRM using n8n and bash scripts. The production units were designed on containers using Docker, allowing for fast and reliable deployment.

Developed a robot that can follow instructions provided by a user using ML algorithms. The robot can navigate in environments that it has never explored before. It is also capable of grasping and manipulating household objects.
The algorithms are based on the transformer, a powerful and novel paradigm for deep neural networks. They were trained on a newly collected dataset on Airbnb and supervised learning.