Xavier Geerinck
Sovereign Data & Agentic AI: From Architecture to Autonomous Systems
Xavier Geerinck is an enterprise software architect, technology entrepreneur, and the Co-Founder & Chief Technology Officer (CTO) of Scrydon, a European sovereign AI and data platform. His work centers on solving the enterprise data sovereignty paradox: enabling organizations in regulated sectors, public entities, and enterprise environments to harness state-of-the-art agentic AI while maintaining complete data ownership, verifiable trust, and strict compliance with European regulatory standards, including the EU AI Act and NIS2.
Across more than 18 years in software and systems engineering, Xavier has worked across the full depth of the technical stack—from low-level networking and distributed, high-throughput streaming systems to semantic knowledge graphs, GraphRAG, and real-time autonomous systems.
Prior to co-founding Scrydon, Xavier founded Composabl (later AMESA), where he pioneered autonomous AI agent training using reinforcement learning and machine teaching for complex industrial automation. Earlier in his career, he operated as an enterprise cloud and software architect at Microsoft, Cisco, and Nokia, partnering with Fortune 500 enterprises, public institutions, and telecommunications providers to design, scale, and secure mission-critical cloud backbones and data platforms.
Alongside his executive role at Scrydon, Xavier is an active researcher, open-source contributor, and technical writer. He authors long-form architectural breakdowns on distributed systems, local LLM deployment, and AI security at xaviergeerinck.com, and creates interactive research tools such as PaperLens (paperlens.io) to make foundational AI research accessible and interactive.
My expertise
My expertise spans the intersection of sovereign data architectures, enterprise AI systems, and high-performance distributed engineering:
Sovereign Data Infrastructure & Federated Data Spaces:
Architecting vendor-independent data fabrics and distributed data spaces that enforce strict data residency, privacy, and mathematically verifiable compliance (EU AI Act, NIS2, and eIDAS 2.0).
Enterprise Knowledge Graphs, Ontologies & GraphRAG:
Grounding non-deterministic large language models into deterministic business reality through formal enterprise ontologies, semantic retrieval (GraphRAG), and end-to-end data lineage.
Agentic AI & Autonomous Systems:
Designing resilient, multi-agent architectures that move beyond simple prompt chains toward mission-critical enterprise workflows, verifiable decision-making, and autonomous industrial control (rooted in reinforcement learning and machine teaching).
Modern Data Platforms & Analytical Engines:
Engineering high-throughput, low-latency data pipelines and modern lakehouse patterns (utilizing technologies such as Databricks, DuckDB, StarRocks, Polars, and Redpanda Connect) optimized for execution speed and architectural efficiency.
Cloud-Native Systems Engineering & Distributed Architectures:
Implementing scalable, air-gapped, and hybrid environments using Kubernetes, Dapr, TypeScript/Next.js, and microservices, underpinned by deep networking foundations (from kernel-level paradigms to zero-trust edge infrastructure).
Sovereign Data Infrastructure & Federated Data Spaces:
Architecting vendor-independent data fabrics and distributed data spaces that enforce strict data residency, privacy, and mathematically verifiable compliance (EU AI Act, NIS2, and eIDAS 2.0).
Enterprise Knowledge Graphs, Ontologies & GraphRAG:
Grounding non-deterministic large language models into deterministic business reality through formal enterprise ontologies, semantic retrieval (GraphRAG), and end-to-end data lineage.
Agentic AI & Autonomous Systems:
Designing resilient, multi-agent architectures that move beyond simple prompt chains toward mission-critical enterprise workflows, verifiable decision-making, and autonomous industrial control (rooted in reinforcement learning and machine teaching).
Modern Data Platforms & Analytical Engines:
Engineering high-throughput, low-latency data pipelines and modern lakehouse patterns (utilizing technologies such as Databricks, DuckDB, StarRocks, Polars, and Redpanda Connect) optimized for execution speed and architectural efficiency.
Cloud-Native Systems Engineering & Distributed Architectures:
Implementing scalable, air-gapped, and hybrid environments using Kubernetes, Dapr, TypeScript/Next.js, and microservices, underpinned by deep networking foundations (from kernel-level paradigms to zero-trust edge infrastructure).
Domain/sector
- Government & Administration
- Computer Science & IT
- Across all disciplines (AI in general)
Organisation
