Applied Artificial Intelligence

In an environment of increasingly sophisticated cyberthreats, Artificial Intelligence becomes a decisive factor to anticipate, detect, and respond to incidents in real time. We develop systems that combine User and Entity Behavior Analytics (UEBA), process mining, and deep learning to identify anomalous activity before it has any impact. We apply AI to fraud prevention, intrusion detection, and automated incident response, strengthening the digital resilience of critical sectors such as finance, healthcare, and defense. Additionally, we integrate Privacy-Enhancing Technologies (PETs) and privacy-preserving machine learning (PPML) to ensure that sensitive data can be processed securely without compromising individual identities.This approach enables proactive and autonomous cybersecurity, capable of anticipating risks, dynamically adapting to new threats, and reinforcing digital trust.

The centralization of data in the cloud poses growing challenges in privacy, efficiency, and sustainability. At Gradiant, we research and develop decentralized AI solutions, where processing is performed at the network edge or directly on devices. Through federated learning and confidential computing, we enable multiple nodes to collaboratively train global models without transferring sensitive data. This reduces both latency and bandwidth consumption. The same philosophy drives our research in AI for communications, where we optimize spectrum usage, enhance signal quality, and ensure resilience in dynamic environments. Our solutions include cognitive networks that autonomously adjust operational parameters and deep learning algorithms that detect interference, predict traffic, and allocate resources optimally. These advances are critical in 5G, 6G, IoT, and satellite platforms, enabling smart, secure, and sovereign connectivity aligned with Europe’s digital sovereignty goals.

The next generation of intelligent systems requires technologies capable of understanding and processing complex, multimodal information. We develop AI that integrates text, image, audio, and video to generate more complete, reliable, and contextual analyses. These techniques are applied in cybersecurity, defense, healthcare, and industry, where a single data channel is insufficient to achieve a comprehensive view. We also explore neurosymbolic AI, which combines machine learning with logical reasoning to create systems that are more interpretable, adaptive, and capable of applying common sense and structured knowledge in decision-making. Finally, we research neuromorphic computing, inspired by the human brain, to bring advanced intelligence to the edge. Using spiking neural networks (SNNs) and specialized hardware, we enable platforms with low power consumption, high parallel processing, and real-time response, with applications in autonomous mobility, industrial IoT, and environmental sensing. These emerging technologies position Gradiant at the forefront of next-generation autonomous and cognitive systems development.