The posting
Why Keyrus, Why Now!
Keyrus is an international group of 2,800 consultants and experts across 28 countries, built on a single conviction: AI does not transform businesses. Architected intelligence does. For more than 30 years, we have been building the data foundations that make intelligent systems work — designing the Operating System of the intelligent enterprise, where intelligence is embedded into the core of business processes to create sustainable value: we operationalise intelligence.
AI does not replace humans. It repositions us to a place no system can follow: understanding, deciding, designing, and creating. At Keyrus, you will not just develop skills - you will develop judgment. Your expertise sharpens with every system you architect, every client challenge you solve, every deployment that compounds on the last.
Over time, you grow into one of the rarest professionals of the intelligence era: someone who bridges data, AI, and human decision-making at scale, across industries and geographies. This is not a role you fill. It is a discipline you master and a story you help write to become a Keyrus Architect of Intelligence.
Technology amplifies. Keyrus culture differentiates. Industrial discipline connects the two.
What You'll Architect
As an AI Foundation Architect, you will define the technical architecture that enables production AI and agentic systems to operate reliably at enterprise scale.
You will define how models, agents, data, applications, APIs, cloud services, security controls, and operational capabilities fit together, and set the architecture principles and boundaries that guide AI Engineers, Software Engineers, Data Engineers, and other technical teams. Within your assigned architecture scope, you are accountable for the integrity of the target design. Depending on the engagement, this may include formal design approval; explicitly define broader enterprise governance authority rather than assume it.
Your primary focus is client-facing architecture and technical design. You will also turn successful approaches into reusable Keyrus architecture patterns and standards without automatically owning an internal platform product or roadmap.
Your responsibilities
- Define target architectures for production GenAI and agentic systems, connecting models, agents, enterprise data, applications, APIs, cloud infrastructure, security, and operational services.
- Establish architecture principles, reference patterns, technical boundaries, and non-functional requirements that engineering and delivery teams can apply consistently.
- Make and document material architecture decisions across platform selection, model access, integration patterns, agent execution, data access, deployment, scalability, resilience, observability, and architecture-level cost.
- Review solution and engineering designs for architectural integrity, approving material design decisions within the assigned scope where the role has been designated as the architecture authority.
- Design secure, scalable, and operable foundations for intelligent systems, ensuring that reliability, identity, access, auditability, recovery, maintainability, and human-control requirements are considered from the start.
- Guide specialist engineering teams without replacing their ownership, working with AI Engineers, Software Engineers, Data Engineers, Governance specialists, and delivery teams on decisions that cross technical domains.
- Engage with client architecture and technical stakeholders, explaining architecture choices, dependencies, risks, trade-offs, and required decisions in clear technical and business terms.
- Convert successful architecture decisions into reusable Keyrus assets, including reference architectures, decision records, standards, patterns, and design guidance that can strengthen future engagements.
- Architecture-level AI FinOps, including model/provider routing, consumption patterns, capacity decisions, and cost optimisation.
Who You Are
You are an experienced technical architect who combines strategic architecture judgment with recent practical understanding of production AI systems. You can move between the complete system and its individual components without trying to become the implementation owner for every layer, and you know when an architecture decision requires depth and when it requires coordination across specialists.
You remain close enough to modern GenAI and agentic engineering to understand how these systems actually behave in production – not only how they appear on an architecture diagram. You are credible with engineers and client architects alike and can explain why an architecture decision matters for security, reliability, scalability, performance, cost, and long-term operability.
What You Bring
Must Have
- Seniority threshold -8+ years designing complex architectures, set locally
- Strong experience designing and owning architecture for complex production technology environments, with clear accountability for cross-system technical decisions.
- Recent practical experience architecting production GenAI, LLM, or agentic systems, including the interaction between models, data, applications, integrations, and operational infrastructure.
- Deep architecture expertise in at least one major cloud and AI ecosystem, with sufficient breadth to evaluate comparable technologies and integration approaches.
- Strong understanding of distributed systems, APIs, integration architecture, identity and access, security, scalability, resilience, observability, and production operations.
- Ability to define and communicate reference architectures, technical standards, interface boundaries, non-functional requirements, and architecture decisions.
- Strong technical judgment and the ability to make transparent trade-offs across performance, reliability, security, maintainability, latency, cost, and operational complexity.
- Data architecture in scope - Lakehouse, data mesh, streaming, feature stores
Should Have
- Experience architecting RAG, agentic, or multi-model systems and understanding the architectural implications of retrieval, model routing, tool access, state, memory, agent execution, and human oversight.
- Experience designing architecture for model access and deployment patterns across managed AI services, APIs, open-source models, or comparable enterprise AI environments.
- Understanding of MLOps / LLMOps, CI/CD, containerized environments, infrastructure automation, monitoring, and lifecycle management from an architecture perspective.
- Experience working directly with client architects, engineering leaders, security teams, and technical decision-makers.
- Experience conducting architecture reviews and guiding multiple engineering disciplines within a shared target architecture.
- Ability to create architecture documentation such as reference architectures, architecture decision records, interface definitions, diagrams, standards, and technical roadmaps.
Could Have
- Experience with AI Agent Platform Engineering, including reusable agent execution environments, tool gateways, permissions, shared observability, or platform-level runtime controls.
- Experience producing reference implementations or technical spikes to validate critical architecture decisions.
- Experience with enterprise architecture frameworks or formal architecture-governance processes.
- Experience with hybrid, sovereign, regulated, or security-sensitive AI environments.
- Experience with complex event-driven, streaming, distributed, or high-scale architectures.
- Consulting or professional-services experience.
Illustrative technology environments – not all required
Depending on the engagement, the architecture may involve cloud and AI ecosystems such as Azure and Azure AI, AWS and Bedrock, GCP and Vertex AI, OpenAI, Anthropic, Google models, open-source model ecosystems, modern data platforms, Kubernetes, API and event platforms, vector or hybrid search services, and enterprise identity and observability solutions. Mistral, Lakehouse, Data mesh, Data fabric, Palantir Foundry, AIP, Kafka, Feature stores, Model serving, Terraform, MCP, AI gateway, Snowflake, Databricks, Apache Iceberg, OpenTelemetry
Candidates are not expected to have equal depth across all of these technologies. The expectation is deep architecture expertise in at least one relevant ecosystem, strong production judgment, and the ability to evaluate comparable technologies based on the requirements of the system.
What Makes You Successful
- You make architecture decisions explicit rather than allowing them to emerge accidentally during implementation.
- You balance technical ambition with production reality, challenge unnecessary complexity, and choose patterns that teams can build, operate, and evolve.
- You know which decisions require architecture authority and which should remain with specialist engineers.
- You use production constraints – security, reliability, latency, cost, scalability, and maintainability – to guide design.
- You communicate architectural trade-offs clearly to both technical teams and client stakeholders.
- You create reusable architecture knowledge rather than solving every engagement from zero.
Rewards - What We Offer at Keyrus
Growth
At Keyrus we have created an environment where people are supported, challenged and connected to the people around them.
Working on impactful Data & Digital projects you will develop a broad and versatile skill set and stay ahead in a fast-evolving, AI-driven landscape.
Our digital learning platform, with over 30,000 courses including in-house trainings developed by our experts, is a perfect way to stay on top of your game
A generous training budget is available to help foster a culture where learning, feedback and continuous improvement is the norm You will have a dedicated coach, benefit from regular check-ins, learn via different communities and have clear alignment with your team and company goals We want you to grow with confidence. We therefore place a strong emphasis on ownership, open feedback and continuous development.
When people feel trusted, connected and inspired, they do their best work!
Rewards
Mobility budget? Car with fuel or charging car? We offer both so you can choose!
On top of that we also have daily and monthly net allowances, eco vouchers, hospitalisation (DKV) insurance and group insurance, as well as a generous bonus scheme that you can use on our flexible reward plan.
What We Stand For
Collective Intelligence — Collaboration across expertise, functions, and geographies makes it possible to combine know-how and deliver more comprehensive responses to client challenges.
Reliability — The ability to deliver complex projects with rigour is one of the pillars of our client relationship.
Pragmatism — Prioritising concrete impact and measurable value over abstract technological discourse.
Entrepreneurial Spirit — Curiosity, energy, and freedom are the foundations of our culture; they enable initiative and sustained innovation.
Our Hiring Process
We keep our process straightforward and respect candidates' time. The local process should normally include:
- A conversation with the Talent team to discuss the candidate's background, expectations, and what Keyrus can offer.
- A technical architecture conversation focused on recent production AI architecture, architecture decisions, system boundaries, non-functional requirements, and the candidate's primary technology ecosystem.
- An architecture case or representative challenge to understand how the candidate evaluates trade-offs, defines a target architecture, and communicates decisions to engineering and client stakeholders, followed by a final conversation with the relevant technical or business leadership. The assessment should focus on architectural proficiency and decision-making rather than the number of technologies a candidate can list.
About Keyrus
At Keyrus, we help organizations move from experimental AI to industrialized AI, from isolated agents to orchestrated systems, and from insight to execution. This is the discipline we call being an Architect of Intelligence. Designing the Operating System of the intelligent enterprise, where intelligence is embedded into the core of business processes to create sustainable value: we operationalize intelligence.
Powered by our proprietary Human Orchestrated Model™ (HOM), we architect reliable:
· Intelligence Foundations,
· Human in Command governance,
· and Performance Steering,
To create intelligent environments where technology amplifies human capabilities and performance compounds over time.
With 30+ years of expertise and 2,800 employees across 28 countries, we help organizations go beyond transformation: to build adaptive, resilient, and continuously improving intelligent organizations.
AI does not transform businesses. Architected intelligence does.
#LI-JA1
