Expertise covers the following technical domains (non-exhaustive) 

Software: cloud computing, edge computing, virtualization, distributed systems, high-performance computing.

Artificial intelligence: Neural Networks (NN), Graph Neural Networks (GNNs), transformers, language models, Generative Adversarial Networks (GANs), diffusion models, mixture of experts (MoE), machine learning, reinforcement learning, federated learning, adversarial machine learning, zero-shot learning, few-shot learning, reinforcement learning from human feedback (RLHF), computer vision, Natural Language Processing (NLP), retrieval-augmented generation (RAG), prompt engineering, chain-of-thought (CoT), genetic algorithms, generative AI, agentic AI, autonomous AI, frugal AI, explainable AI (XAI), recursive self-improvement... 

Quantum technologies: quantum key distribution, quantum sensing, post-quantum cryptography, quantum error correction.

Cybersecurity: zero-knowledge proofs, homomorphic encryption, secure multiparty computation, intrusion detection.

Blockchains & distributed ledgers: smart contracts, DeFi, tokenization, consensus, layer-2 scaling, cross-chain bridges, DAOs.

Telecommunications: next-gen networks, IoT, mesh networks, GNSS, signal processing.

Aeronautics & space: drones, swarms, satellites, avionics, flight management systems.

Medical devices: diagnostics, wearables, sensors, surgery, telemedicine, medical imaging, implantable devices.

Robotics: humanoid robots, autonomous systems, mechatronics, actuators.

User interfaces: VR, AR, MR, spatial computing, haptics, brain-computer interfaces, eye tracking, HUD.

General engineering & exploratory subjects

François Veltz handles patent drafting in general mechanics and physics. He is also available for exploratory or niche subject matters requiring specific compliance frameworks (e.g. ethical AI alignment, Charia-compatible algorithms).

Exclusions

François Veltz does not operate in the chemical, biological, or pharmaceutical sectors. For these fields, clients are referred to specialized partners.