Post Doc

**About the role**:
Your role is to carry out a post-doc on generative models applied to mobile networks.

**Background and problem**

The amount of location data generated by mobile devices has increased dramatically over the past 20 years; this is the case for portable devices (such as smartphones) and electronic objects (such as smart watches). Furthermore, mobile objects such as cars, trains, buses, etc. are also generating events on the networks. Many economic players are interested in using statistical analysis from these data sources (tourism, transport, geomarketing, logistics...). Public statistics can also be improved by using these new sources.

Data privacy regulations impose constraints on the processing of personal data (European GDPR and ePrivacy directives). To calculate the statistics required by the private sector and public statistics bodies, acceptable legal scenarios must be proposed (use of pseudonymized personal data, retention period before deletion or anonymization, dedicated aggregation algorithms with proven irreversible anonymization...). Respecting existing legal constraints, Orange has developed a commercial solution called Flux Vision to market anonymous statistics based on the analysis of cell phone traces.

In this context, the aim of the post doc is to explore deep generative models to achieve synthetic traces with a broad spectrum of uses.

**Scientific objective - results and challenges**

The aim of this post-doc is to propose and implement generative models applied to the field of mobile network connections, to achieve the generation of synthetic longitudinal traces. These generative models, based on deep learning techniques, will be applied to pseudonymized location histories. The proposed solution will also need to incorporate privacy protection mechanisms to prevent reverse engineering of the traces used to build the model.

**About you**:
**Skills** and personal qualities required by the position:

- Deep learning techniques
- Deep generative models
- Differentially Private Data Generative Models
- Mobile signaling data processing

**Programming & Algorithms**: a high level of proficiency in one of the deep learning programming languages (e.g. Python, etc.) is required.

**Soft Skills**: Initiative, scientific curiosity, self-starter.

**Education required** : PhD in computer science or data science

**Desired experience** : Experience in Deep Learning, generative models

**Additional information**:
The algorithms developed during this post-doc may be implemented on real data to reinforce Orange Business' Flux Vision product line.

**Department**:
Description of the entity to be copied in My job and completed with the description of the team

Orange Innovation brings together the research and innovation activities and expertise of the Group's entities and countries. We work every day to ensure that Orange is recognized as an innovative operator by its customers and we create value for the Group and the Brand in each of our projects. With 740 researchers, thousands of marketers, developers, designers and data analysts, it is the expertise of our 6,000 employees that fuels this ambition every day.

Orange Innovation anticipates technological breakthroughs and supports the Group's countries and entities in making the best technological choices to meet the needs of our consumer and business customers.

Within Orange, you will be part of a research team of about 20 engineers and researchers, mainly data scientists, working in the field of mobility in cellular networks and on advanced processing and anonymization methods, with the aim of responding to societal or industrial issues. This team is at the root of the 'Flux Vision' service, which aims at producing large-scale statistics on the mobility of people and objects based on mobile traces. The team is based at Orange Innovation in Belfort.

**Contract**:
Post Doc