En poursuivant votre navigation, vous acceptez l'utilisation de cookies destinés à améliorer la performance de ce site et à vous proposer des services et contenus personnalisés.


Data scientist

Who are we?

Subsidiary of the Ecole polytechnique, we are FX-Conseil, a private service company specialized in partnerships with research centers. We support the emergence and development of technological transformation, products and services thanks to the expertise of research centers, in particular the research center of the Ecole polytechnique. We connect big companies, SMEs and start-ups willing to benefit from our partnerships with research laboratories simply and easily.

We are starting a new research partnership with a world leader of the industry of electronic health-care databases. This is a project with a huge potential, involving the development of big-data techniques and machine learning algorithms for health. The database records every health-related transaction for millions of people.

We are looking for a Data Scientist to work on this new project: she/he will work on the development of the machine learning pipeline applied on this electronic health record database (from one of the world’s main actors of this industry).

Our team is composed of several enthusiast researchers (international researchers in machine learning from Ecole polytechnique) and data-driven software engineers / data-scientists.

What we want to do

In order to extract knowledge from the database, we want to design and implement new machine learning algorithms. Every good idea can be implemented and tested in a homemade machine learning library, dedicated to the project. We are used to do this using C++/Python, but are open to anything else. The main workflow will look like this:

  • Cleaning: from a production SQL database, we format the data in a large denormalized table;
  • Featuring: we compute a lot (a lot) of features based on this table to create a large matrix;
  • Learning: we feed this preprocessed data to homemade machine learning algorithms.

Typical day

  • You hear about a new overpowered algorithm. You decide to try it out / reimplement it in the pipeline. You ensure that everyone can benefit from it and easily run your awesome code on his machine. You do this in a clean way, through a git workflow;
  • You can use large machines to run your code;
  • You are also part of every decision regarding the evolution of the pipeline’s design and API and can suggest new algorithms to be implemented in it;
  • You walk across the campus and cross some students, teachers and horses;
  • You can use the sport facilities on campus if you want, including swimming pool, sky diving or golf, among many others;
  • When something is unclear about the data, or when you feel the need to, you go and spend the day at our partner’s office, located in a brand new comfy tower in La Defense. The data will be there, but there’ll be a remote secure access to work from FX Conseil.
  • You are also looking for new libraries, releases, papers, conferences, literally anything that’s linked to the field of interest.

Who are you?

  • Curious is your middle name. Github is your facebook. Python is your mother tongue;
  • You spent several years studying computer science and applied mathematics and played with data-oriented tools and machine learning libraries. You acquired good programming skills and know about clean code, tests and versioning;
  • You’ve seen things. Mostly weird implementations and hairy publications. You already applied all of this knowledge for some years in a data-centric startup or group;
  • Also, you know about the machine learning ecosystem and everything revolving around data processing.


You are either:

  • A killer in Python; Fluent in the python machine learning stack and libraries;
  • Know also about R;
  • Not scared to implement new algorithms

You will:

  • Work in a highly challenging intellectual environment;
  • Benefit from the campus infrastructure, which means that you could practice horse riding, golfing and water-polo, unfortunately not at the same time;
  • Use a lot of different algorithms and technologies and try any that looks appealing to you.
  • Discover that machine learning is not only about logistic regression, boosting or deep nets

Salary / type of contract

  • Fixed-term contract
  • Competitive salary


Sylvie Tonda-Goldstein

Nous contacter