AI Architect

Mission Overview

We are seeking an AI Architect for a consultancy mission at a client site, representing Keystone Solutions. The AI Architect will define and oversee the target architecture, reference architectures, and integration patterns for AI. This role ensures that models, agents, data flows, gateways, platform services, and applications coherently align with the broader architecture and meet security, privacy, continuity, and manageability requirements.

Responsibilities

• Develop and maintain the target architecture for generative AI, machine learning, and agentic AI.
• Establish reference architectures and decision patterns for on-premises, cloud, and hybrid deployments, tailored to sensitivity, integration needs, and operational requirements.
• Define architectural frameworks for AI Gateway, Model Context Protocol, retrieval, model access, agents, identity, logging, observability, and lifecycle management.
• Guide delivery teams in architectural choices, trade-offs, and the transition from sandbox or PoC to production in collaboration with technology architects.
• Conduct architecture reviews and timely highlight deviations, technical debt, dependencies, and risks.
• Integrate privacy, least privilege, data minimization, isolation, and traceability by design in collaboration with security and data functions.
• Identify and prioritize reusable platform and integration building blocks with the AI Enabling Team and the Internal Developer Platform.
• Translate technological evolutions into concrete impacts, options, and decision points for the client.

Technical Competencies

• Extensive experience in solution architecture or IT architecture in enterprise environments.
• Experience with APIs, event, and integration patterns.
• Understanding of LLM and AI platform architectures, retrieval, model serving, agents, evaluation, and observability.
• Ability to translate regulations and security requirements into practical architectural principles and guardrails.
• Experience with architectural decision records, reference architectures, non-functional requirements, and lifecycle management.
• Knowledge of the public sector, G-Cloud, or OISZ context is a plus.

Generic Competencies

• Innovation: You keep up with the latest trends and experiment with innovative AI technologies.
• Problem-solving: You analyze complex issues and translate them into pragmatic solutions.
• Collaboration: You work smoothly with colleagues from various disciplines (IT, business, security, architecture).
• Communication: You can clearly explain technical concepts to both technical and non-technical stakeholders.
• Advising: You excel in structuring, visualizing, advising, and facilitating multidisciplinary decision-making.
• Result-oriented: You monitor deadlines and ensure quality deliverables.

Selection

Please attach a document of no more than 2 pages to your application, outlining your vision on how the Federal Pension Service can safely and responsibly leverage artificial intelligence to create tangible value for its pension services and internal operations. Emphasize a coherent, scalable, and manageable AI architecture and the contribution you can make as an AI Architect in defining the target architecture, reference architectures, and technical frameworks, integrating AI solutions into the existing IT landscape, and guiding the transition from experiment to production.

After receiving the CVs, they will be analyzed, and a shortlist of 3-5 candidates will be created based on the criteria established in the ProUnity contract. These candidates will then be invited for an interview where their technical and generic competencies, as well as their language skills, will be evaluated. Subsequently, a ranking will be established.

Work Location: The position is located in Brussels. We offer the possibility of part-time teleworking.

Duration: 14/12/2026 - 13/12/2027 12 months • (full time)

Skills required

• API-, event- en integratiearchitecturen en -patronenConfirmé
• Architectural Decision Records, referentiearchitecturen, non-functionele vereisten en lifecyclebeheeExpert
• LLM- en AI-platformarchitecturen, inclusief retrieval, model serving, agents, evaluatie en observabiJunior
• Solution- of IT-architectuur in complexe enterpriseomgevingenExpert
• Vertaling van regelgeving, privacy- en securityvereisten naar architectuurprincipes en guardrailsConfirmé

Language requirements

• NéerlandaisConnaissance active
• AnglaisConnaissance passive
• FrançaisConnaissance active