28 May 2026

Agentic AI Domain Architect

Carollo Engineers

Amsterdam

28 May 2026

Apply

We are looking for a domain architect to shape the future of AI-powered, agentic-based workflow experiences across logistics, fleet, and mobility solutions. The role sits at the intersection of domain architecture, agent design, product development, and applied data science. Requires 5+ years of experience designing workflow-centric systems and strong background in domain or enterprise architecture.

HybridFull-timeSeniorAPIsEvents

Agentic AI Domain Architect

Apply

HybridFull-timeSeniorAPIs

Salary

Not specified

Work Location

Amsterdam, North Holland, Netherlands, NL

Work Model

Hybrid

Experience Required

5 years

Employment Type

Full-time

Experience Level

Not Applicable

Core Qualifications

Technical (Must-have)

APIseventsMCP-style tool interfacesschemascontractsLLMsagent frameworksERPTMSFSMdata platformsevent-driven architecturessemantic models

Soft Skills

communicationtechnical depthbusiness contextoperational contextcross-functional collaborationworking in ambiguitydelivery-orientedarchitecture-first thinking

Key Responsibilities

•Design and own agentic integration patterns between customer applications, enterprise systems, and agent frameworks (APIs, events, MCP-style tool interfaces, schemas, contracts).
•Define how agents invoke, coordinate with, and reason over external systems while respecting customer architecture, security, and governance constraints.
•Act as the architectural authority on how agentic capabilities are embedded into real customer environments, not as standalone copilots.
•Translate end‑to‑end business workflows into agent-compatible domain workflows, decomposed into modular, reusable agent skills.
•Define skill boundaries, preconditions, outputs, confidence signals, and failure modes.
•Ensure workflows support automation, human-in-the-loop, escalation, and explainability by design.
•Align domain workflows with multi-agent or hierarchical agent orchestration models where needed.
•Design domain-specific context models that combine operational data, spatial/temporal state (where relevant), user intent, and historical interactions.
•Define what agents should remember, forget, summarize, or abstract over time (short-term, long-term, episodic memory).
•Drive a data-centric approach to agent improvement by extracting signals from agent memory, user interactions, workflow outcomes, corrections, and overrides.

Agentic AIDomain ArchitectWorkflow DesignAILogisticsFleetMobilityHybridFull-timeSenior

Core Qualifications

Technical (Must-have)

APIseventsMCP-style tool interfacesschemascontractsLLMsagent frameworksERPTMSFSMdata platformsevent-driven architecturessemantic models

Soft Skills

communicationtechnical depthbusiness contextoperational contextcross-functional collaborationworking in ambiguitydelivery-orientedarchitecture-first thinking

Key Responsibilities

•Design and own agentic integration patterns between customer applications, enterprise systems, and agent frameworks (APIs, events, MCP-style tool interfaces, schemas, contracts).

•Define how agents invoke, coordinate with, and reason over external systems while respecting customer architecture, security, and governance constraints.

•Act as the architectural authority on how agentic capabilities are embedded into real customer environments, not as standalone copilots.

•Translate end‑to‑end business workflows into agent-compatible domain workflows, decomposed into modular, reusable agent skills.

•Define skill boundaries, preconditions, outputs, confidence signals, and failure modes.

•Ensure workflows support automation, human-in-the-loop, escalation, and explainability by design.

•Align domain workflows with multi-agent or hierarchical agent orchestration models where needed.

•Design domain-specific context models that combine operational data, spatial/temporal state (where relevant), user intent, and historical interactions.

•Define what agents should remember, forget, summarize, or abstract over time (short-term, long-term, episodic memory).

•Drive a data-centric approach to agent improvement by extracting signals from agent memory, user interactions, workflow outcomes, corrections, and overrides.