Antoine Gallix

I provided support for deploying machine learning algorithms in production by developing and maintaining a library of tools and framework for the data science team. The main working environment was Google Cloud Platform. Monitoring and predictive algorithms were packed in Docker containers and integrated with the rest of Infarm logistic and remote control platform with the database and process communication tools of GCP.

Highlights

• Develop and maintain a library of functions and CLI tools to integrate with the databases
• Design a framework to deploy the algorithms in containers
• Schedule, operate and monitor the data solutions
• Tech Stack: Python, Pandas, Google Cloud Platform, Docker, SQL

Infarm is an indoor farming company that was in rapid growth, with a rapidly evolving operation model. The software team developed an internal application to coordinate the seeding, transplanting and harvesting operation on the company's distributed growing machine network. I was in charge of the backend part of the logistic service.

Highlights

• Designing and planning data model evolution
• Developping and coordinating new features with the frontend team.
• Maintenance, Bug Fixing
• Tech Stack: Python, Flask, Postgresql, Heroku, Github

CeleraOne started a new branch of business, that uses machine learning technologies to detect defects in the steel production industry. A new team was gathered and a pilot project was run for half a year.

Highlights

• Exploratory analysis of the dataset
• Researching, implementing and evaluating different approaches
• Complex data processing, development of a common toolbox for research
• Tech Stack: Python, Pandas, Matplotlib, SciKitLearn

CeleraOne provided services for online publishing industry; various types of content access metrics and paywall functionalities. I was part of a team developing and maintaining the backend of the web application used by customers to configure the system.

Highlights

• Development and maintenance of a backend and REST API
• Promoted good clean code practice, refactoring and documentation of a codebase with a heavy technical debt.
• Tech Stack: Python, Pyramid, Colander, MongoDB, Buildout, Git

In an early stage startup, I started building the backend of a web application that manages agricultural data.

Highlights

• Tech Stack: Python, Eve, Flask, MongoDB

One of the products of DxO is image processing algorithms, that work on-chip on smartphones, to create the image from raw sensor data in real time. I worked on a comparative study of various algorithms on one of the steps of the image pipelinen.

Highlights

• Tech Stack: Python, C++, Mercurial

Master Thesis Project. Application of image decomposition and machine learning techniques to classification of 3D brain scanner images

Highlights

• Publication of the research in a peer reviewed journal

Second Year of Master Studies in Signal and Image processing

First Year of Master's Degree

The subject of the study is automatic classification of brain SPECT images, in the case of Alzheimer disease. In this paper a new step of frequency decomposition ( Empirical Mode Decomposition ) is applied to an existing classification algorithm that leads to an improvement of the results.