Supervising 3 teams (12 engineers) : mobile mapping sensor, Web & GIS software development, and Research.Overseeing interdisciplinary projects to strengthen the commercial range of Imajbox sensors, Imajview geographic data production software, Imajnet web data distribution platform, and processing chains for automated asset inventory.Implemented Agile methodologies to continuously align customer needs with development priorities. Developed best practices and automated workflows for machine learning projects to accelerate model production.

Supervised a 4-engineer R&D team in computer vision & machine learning. Defined and conducted the research strategy to support the company vision.Coordinated R&D projects (€750k / 3 years) to develop automated asset inventory products for linear infrastructure (achieving 94% recall/precision) using object detection, classification and clustering.Designed and developed navigation algorithms (IMU/GNSS/vision), tracking of detected assets (computer vision), inventory reconciliation (optimization), and change detection (statistics).

Involved in challenging missions for various clients including CAC 40-listed companies and European administrations:- Design of computationally efficient algorithms for geospatial data processing chains to reduce cost on cloud-based infrastructures.- Benchmark of market graph databases to support a data platform dedicated to intelligence missions.- Classification of 4000+ software applications using IT data to identify critical and obsolete ones. - Design of a complete training on the data landscape (quality, security, modeling, governance...) to make our clients more proficient on this subject.- Analysis of complex financial data to value a customer portfolio (joint venture creation).

Participated in the Space Systems Engineering Advanced Master with a short course on the subject "Satellite localization for animal tracking: radio and underwater positioning with Argos"​

Designed and developed algorithms for tracking underwater mobiles (GNSS/light/magnetic field/pressure sensor fusion) and for segmenting fishing ship trajectory (clustering). Published results in scientific journals and international conferences.Conducted and managed performance studies for satellite telecommunications systems (€355k / 2 years). Designed and developed performance measurement software. Carried out statistical analysis of acquired data.

Designed and implemented a new localization algorithm for the 20,000 platforms of the ARGOS system. Reduced positioning error by a factor of 3 and increased the number of delivered positions (up to 2x) for platforms operating under challenging conditions.Improved existing algorithms in the fields of multi-model filtering (extension to heterogenous state-space models) and smoothing (reduced complexity). Published results in scientific journals and international conferences.

Designed a proof of concept to improve ARGOS Doppler geolocalization algorithm.