11/01/2026 updated
AI Architect & Fractional CTO / CAIO | LLM & Multi-Agent Systems | NLP & ML Expert
About me
AI architect and fractional CTO / CAIO. 17 years of professional AI work, PhD in Computational Linguistics (Marie Curie Fellow). I design and secure LLM and agent systems (isolated agent execution, MCP tool governance, RAG, evaluation) and lead them into production. Based in the Netherlands.
AI architect and fractional CTO / CAIO with 26+ years in technology, 17 years of professional AI work (since 2009) and a PhD in Computational Linguistics (Marie Curie Fellow).
AI and agentic systems
LLM pipelines, multi-agent orchestration, retrieval-augmented generation (RAG), MCP and function calling. Agent harness design: context, tools, permissions, isolation, verification, evaluation and guardrails.
Designed and operate a governed multi-agent software delivery platform with isolated execution cells, scoped machine identities, governed tool surfaces, independent merge checks and post-merge verification. More than 14,000 merged pull requests. Built the AI Radar publishing engine.
Production LLM platforms
LangGraph supervisor graphs and Postgres checkpointing, FastAPI, pgvector retrieval and Pydantic validation. Model-agnostic routing across OpenAI, Anthropic, Gemini, Mistral and Cohere, with tenant-level configuration and token/cost tracking.
Cloud-native delivery
Kubernetes on Azure AKS and AWS EKS; Helm, KEDA, Kubernetes Jobs and Key Vault CSI. Terraform for AKS, container registry, API Management, PostgreSQL, Redis, Key Vault, private endpoints and DNS. AWS SageMaker, Lambda, Step Functions, DynamoDB, S3 and EC2; Azure AI; serverless, microservices, IoT data and MLOps.
Observability and release assurance
LangFuse and LangSmith tracing; Prometheus, Grafana and Loki. Nightly LLM evaluations, load and resilience tests, GitHub Actions CI/CD, Playwright end-to-end tests, Trivy scanning and OpenAPI contract checks.
Technical leadership
Fractional CTO / CAIO / CIO: vision, architecture decisions, team formation and investor alignment. Led five teams and more than 30 people across permanent staff and contractors, onshore and offshore. AI readiness assessments, responsible-AI frameworks and compliance-aware design.
NLP, ML and computer vision
Information extraction and retrieval, recommender systems, machine translation and terminology extraction. Computer vision and deep learning on drone, satellite and mobile imagery; production GAN systems.
Languages: English (fluent), Portuguese (native), Spanish (professional working proficiency). 34 publications, more than 30 peer-reviewed.