Pioneers in edge computing and embedded AI
Since AzurİA was founded in 2021, we have been developing innovative embedded AI solutions designed to operate at the edge, as close to the sensors as possible, with a controlled environmental impact.
Our expertise in frugal AI, rooted in applied research, and our technological vision have enabled us to position ourselves as pioneers in real-time embedded processing.
As early as 2019, our founders successfully led a research project in artificial intelligence on a nanosatellite, demonstrating the feasibility of deploying resource-constrained AI algorithms in extreme environments.
This foundational breakthrough now forms the technological backbone of AzurİA and guides the development of robust, efficient, and operational embedded AI solutions dedicated to protecting the planet and its inhabitants.
Edge Computing
Image processing is performed at the edge, as close to the sensors as possible, to extract only relevant information.
This approach significantly reduces data flows, improves real-time responsiveness, and lowers the need for connectivity and cloud infrastructure.
Eco-designed
Our solutions are eco-designed from the outset, with a clear objective: to minimize energy consumption and carbon footprint.
Local processing and a minimal hardware footprint enable controlled energy dissipation, even in constrained environments.
Frugal
Thanks to frugal AI models that can be trained with limited labeled data, our algorithms quickly adapt to new use cases or detection classes, without additional hardware costs.
Multispectral
The integration of complementary spectral bands (visible, infrared, etc.) significantly enhances detection performance, particularly compared to conventional RGB solutions in complex conditions.
Systemic approach
Our models are evaluated not only on their AI metrics, but also on system constraints: inference time, overall power consumption, robustness, and operational integration.
This approach ensures AI that is truly deployable in the field.
Agnostic
Our AI is sensor-agnostic and adapts to different input spectral bands, enabling inference on images from a wide range of sensors, in both the visible and infrared spectrum.
Eco-responsible and ethical by design
Dataset selection
When using external datasets, we conduct a rigorous verification of their origin, collection conditions, and usage licenses.
This approach ensures responsible AI that is compliant with legal frameworks and respectful of data rights.
Aİ training
Images are annotated semi-automatically and then reviewed internally by our data scientists.
Model training is carried out on local GPUs, with continuous monitoring of energy consumption to control the carbon footprint of our AI developments.
Solutions
Inference is performed in real time, close to the sensors, on low-power hardware.
Only the alert containing relevant information (snapshot, metadata, geolocation) is transmitted to the end user, ensuring:
– GDPR compliance,
– A drastic reduction in data flows,
– Minimal environmental impact, by avoiding the transmission and storage of energy-intensive video streams
Frédéric FERESIN
-
University graduate (DEUG S) before joining an engineering school (Polytech Orléans).
-
30 years of experience in engineering and complex project management within major space industry groups, covering phases A through E
-
Led in 2018 the IRT project “Autonomous and Reactive Image Chain (CIAR)”, focused on deploying embedded AI on nanosatellites and ground drones.
-
Co-founder of AzurİA, driven by a clear vision: to harness deeptech, edge computing, and frugal AI in service of protecting the planet, territories, and populations.
Lionel Daniel
-
PhD in Logic, with research in probability theory and biometrics applied to artificial intelligence.
-
4 years of research in logic and probabilistic AI.
-
6 years of software engineering in the field of space-based imaging systems, including optical payloads, in-flight calibration, and image processing.
-
3 years as AI architect on the CIAR project, with a focus on frugal, real-time embedded AI.
-
Responsible for the algorithmic, software, and scientific robustness of the embedded AI solutions developed by AzurİA.
-
Integration of the PACA Est Incubator
-
National i-Lab pre-selection, validating the scientific and industrial relevance of the project.
-
Founding of AzurİA based on strong values embedded in its statutes (Social and Solidarity Economy).
-
Awarded the French Tech Emergence grant to develop frugal, multispectral embedded AI.
-
Winner of Confiance.ai, a program of the French Grand Challenge on artificial intelligence, for the development of low-power embedded AI.
-
Winner of the European SecurIT call for projects, to develop Helia, an intelligent tethered aerostat for real-time detection and alerting.
-
Selected as “AI Talent” in the European BonsApps project, dedicated to the development of artificial intelligence solutions for industry.
-
Signing of the RAPID project, focused on integrating embedded artificial intelligence algorithms into an observation satellite constellation.
-
Selected for a real-world demonstration at the i-Naval event, showcasing an AI-based solution for detecting events at sea.
-
Recipient of the Crédit Agricole Innovation Awards, recognizing high-impact technological solutions.
-
Awarded the Hi France label, recognizing excellence in French artificial intelligence projects.
-
Real-life demonstration of Helia for forest fire detection
-
AMO phase 1 architecture protection of the Sologne forest (SDIS 18/41/45)
-
Phase 2 AMO (project management support) for the deployment of a forest protection architecture in the Sologne region (SDIS 18/41/45).
-
Recipient of the SME Award for commitment to ecological transition and CSR.
-
Winner of the EcoNum project under the France 2030 program, for the development of an eco-designed embedded AI device.
-
Awarded the regional Economic Intelligence label, recognizing strategically innovative companies for regional development.
-
Deployment of an early wildfire detection solution in the Alpes-Maritimes department (06).
-
Deployment of an AzurİA solution on board the Ocean Viking vessel operated by SOS Méditerranée, enabling automated detection of shipwrecked individuals and boats at sea using embedded AI.
-
Deployment of an early wildfire detection solution in Thailand, illustrating the international application of our embedded AI for environmental monitoring.
-
Deployment of solutions in the Loire department for the detection of wildfires and illegal dumping, supporting territorial protection efforts.
-
Launch of AzurİA’s fundraising round to accelerate the industrial deployment of our embedded AI solutions.
They support us
