Netherlands, BARNEVELD
AI Platform & Agentic Architect d/f/m
Location(s):
Essen, NW, DE, 45141
RWE AG
To start as soon as possible, full time / part time, permanent
As an AI Platform & Agentic Architect you design the shared AI platform every use case runs through: a model gateway, an agentic architecture, and guardrails and sovereignty built in from the start. Build it once, make it reusable, keep it safe and vendor-neutral.
About the role - Own and evolve the reference architecture for our AI platform, including the model garden, a governed model gateway providing unified access to all models, and the orchestration of AI agents. - Define enterprise standards for agentic AI, covering approved agent frameworks (commercial and open source), MCP-based integrations, multi-agent coordination patterns, and scalable agent memory architectures. - Set standards for agent evaluation, reliability and observability, so agents are dependable in production, not just in demos. - Design the gateway for reuse, control, audit, provider-independence and cost/token transparency. - Build governance, security, and digital sovereignty into the AI platform, including EU AI Act compliance evidence and audit logging, guardrails, data residency controls, customer-managed encryption keys, and a comprehensive agent control plane covering machine identities, agent registration, and release governance. - Define the paved-road patterns and sandbox-to-production paths that let teams move from idea to value in weeks.
Job requirements and experience - University degree (or equivalent) in Computer Science, Computer Engineering, or Information Technology or related field. - Several years of hands-on working experience in AI platform architecture and agentic AI in a regulated decentral environment. - Deep experience in designing and running AI/ML or GenAI platforms at scale. - Strong agentic AI expertise: agent orchestration and multi-agent frameworks (e.g. Copilot Studio, Databricks,...
Publicēts pirms 4 dienas(-ām)