Xabier Muruaga

I shape production AI and GenAI from business value to deployable architecture, with security, assurance, and auditability built in. The objective is simple: turn promising use cases into systems that can operate reliably at scale.

In large organizations, GenAI stalls for predictable reasons. Prototypes move quickly, but they break at scale when decision boundaries are unclear, controls fragment across teams, and accountability is reconstructed after the fact.

I make scale deliberate by defining reference architectures, operating boundaries, and acceptance criteria teams can actually follow. Security and assurance are built into runtime and release gates, so controls and evidence do not have to be recreated on every deployment.

When systems are already live, I assess them through threat modeling, adversarial testing, and assurance reviews focused on GenAI risks such as prompt injection, data leakage, and unsafe tool or permission boundaries.

The case studies reflect this lifecycle: architecture at scale, AI security architecture, independent assessment, governed agent autonomy, compliance readiness, and value selection.