Why join Artelys?
Joining Artelys means above all joining a dynamic and motivated team, a stimulating work environment and participating in varied and exciting projects. More than 30% of Artelys’ activity is devoted to R&D projects, which allows its engineers to explore state-of-the-art applied mathematical methods in Operational Research, Machine Learning and Deep Learning.
Among the R&D projects that are currently being developed involving the study of advanced Data Science models, we can cite the NEXT project in particular.
As a large-scale R&D project of the French Environment and Energy Management Agency (ADEME), NEXT aims to develop an innovative reference software for the dimensioning of flexible electrical grids (smart-grid), thus providing an essential building block for the energy transition. The core of the NEXT project is the understanding and modelling of the electricity consumption process on a very fine temporal and spatial scale. In particular, modern clustering techniques make it possible to highlight common consumption characteristics. On the other hand, this project is an opportunity to implement and test numerous predictive modelling techniques, from classic GAM models to modern machine learning techniques and in particular Generative Adversial Networks and Variational AutoEncoders.
Partners: Gaz Electricité Grenoble, INRIA, L2EP
Artelys’ Data Science activity is growing rapidly, with numerous client projects, particularly in the Energy, Transport/Mobility, Health and Public Sector (Ministries, Local Authorities, National Agencies) sectors
Some examples of client projects in which a Senior Data Scientist profile may be involved include:
For RTE R&D: Development of a tool (package R) for the study of short-term (D+15) forecasting models for national photovoltaic and wind energy production. This tool must be able to challenge the current models used by RTE, which are models at the scale of a power plant whose forecasts are then aggregated. To do so, Artelys has implemented new “machine learning” models (GAM, XGboost) and exploited new data sources (ECMWF). These models are used in an operational framework with 10-minute granularity forecasts and the integration of “real” measurements taken every hour to adjust the forecasts. As part of this service, probabilistic forecasts by quantile regression of mean forecast errors have been implemented to complement the mean estimate and provide additional risk management tools.
As part of a dynamic, high-level team, you will work jointly on R&D and customer projects, focusing on the analysis and consolidation of data (large and small), the design and implementation of high-performance modelling methods, as well as the visual restitution of these processes.
In particular, you will be brought to:
- Manage customer or internal projects, by providing technical added value to your teams composed of Data Scientist and Data Engineer and by managing the relationship with the customer.
- Participate in the development of technical proposals based on advanced methods of applied mathematics, in order to meet the needs of our customers.
- Exploring raw data sets (potentially very large and from disparate sources), implementing visualisation techniques to intuit their meaning, applying consolidation methods to give them a robust structure as a basis for modelling.
- Implement and test a wide range of modelling solutions (from classic statistics to modern machine learning techniques), in order to exploit the best combination of them.
- Develop solutions for visualising and exploiting results in an operational context.
- Define and implement extensible, scalable architectures
Holder of a Master and/or PhD with a specialisation in the field of data-sciences and statistics, you have at least 3 years of professional experience.
You have a dual competence in mathematics and computer science, particularly scientific (Python and/or R). You have skills in relational databases (PostgreSQL or equivalent) or non-relational databases (Elastic Stack in particular).
You are the ideal candidate if:
- You are curious and you approach the mining of new data sources like a game.
- You master and have implemented clustering, classification and regression methods.
- You consider that the understanding of the business context associated with a data problem is essential to its resolution.
- Driven by a pronounced pragmatism in your modelling choices, you know that the best model is not necessarily the most complicated.
- You are aware that the world of “Small Data” involves as many challenges and issues as “Big Data”.
- You are autonomous, rigorous and driven by a strong team spirit.
In addition, a good fluency in French will be highly appreciated.
Experiencing a strong growth, Artelys is looking for proactive and creative employees, conscious about the added value of their work and motivated to take part in the development of our activities.
Permanent position in our Lyon office.
This is your dream job? Apply now!