Offer Description
Context.
This position is funded by the ANR PRCI LUCAS project. The project aims to leverage external data to represent, analyze, understand, and identify the causes of anomalies observed in water networks. It seeks to move beyond traditional single-source problem analysis approaches and instead harness independently designed databases to better address challenges in urban networks. We will use water networks as a case study to illustrate the generic methodologies developed within this project.
Water network data is rich and diverse, including Geographic Information Systems (GIS) that provide information on network infrastructure, as well as videos from televised inspections (CCTV), which are essential for annotating various anomalies observed in sewer pipes. Our primary goal is to go beyond water network-specific data to include contextual information about the network environment, such as buildings, real-time traffic, population density, etc. This project aims to propose innovative solutions to the challenges posed by these massive and diverse datasets.
This includes collecting and representing data, as well as processing and selecting relevant data for targeted anomaly analysis. Subsequent tasks involve developing effective methods to combine this data to identify, understand, and explain the causes of observed anomalies—and, most importantly, to predict them. Special attention is given to causal attribution, which involves determining which of a series of time-distributed events is responsible for the occurrence of an anomaly. The methods developed must account for the incomplete and uncertain nature of the data, which is intrinsic to water network data analysis.
The project consortium is multidisciplinary, bringing together researchers specializing in water modeling, data science, and artificial intelligence to tackle the identified challenges.
Assigned Mission.
The recruited individual will participate in Work Package 1 (WP1) of the project, which involves collecting and locating data on urban water networks, identifying and selecting key elements, and preparing them for integration into other work packages. Participation in drone-based data collection is also envisioned to capture surface elements that may help identify anomalies not directly visible.
A key task will be ensuring the geolocation of data, which is sometimes provided without spatial references. All data will be converted into a graph format to facilitate matching between different sources, similar to the approach taken in [1]. The use of multi-source data will help understand and attribute the causes of pipe anomalies. Causal inference models (e.g., [2]) will be used to establish relationships between observed anomalies and environmental factors. Finally, the focus will be on developing algorithms and proofs of concept capable of integrating data from GIS, videos/photos, and inspection reports to detect, classify, and attribute the causes of anomalies.
The missions and objectives may be adapted based on the skills of the recruited individual.
Main Activities.
- Programming in Python and/or C++
- Geomatics
Funding.
As the position is funded under the ANR LUCAS project, the recruited individual will be required to report on their work both in writing and orally, in English and French.
The work will be carried out at the IUSTI Laboratory of Aix-Marseille Université (AMU) in collaboration with the CRIL in Lens.
Contact.
For further information and initial contact, please send an email with the subject “LUCAS Postdoc” to carole.delenne@univ-amu.fr, attaching a CV, before June 15, 2026.
Where to apply
E-mail: carole.delenne@univ-amu.fr
Skills/Qualifications
Required Skills.
- Proficiency in Python, particularly for geographic data.
- Image/signal processing
- Knowledge of hydraulic models would be a plus.
Ideally, the recruited individual will hold a PhD in Computer Science with a strong command of the Python programming language.
