DevOps / AI Infrastructure Engineer for Agentic Systems

About the role

We are looking for a DevOps / AI Infrastructure Engineer to design, build, and operate the platform that powers our AI‑driven finance operating system. You will work closely with AI leadership and engineering teams to turn experimental LLM workloads into reliable, auditable services for Swiss SMEs and fiduciary partners.

Key responsibilities

• Design and manage cloud infrastructure for AI‑powered services on GCP (Zurich multi‑zone) using Terraform and Kubernetes.
• Build CI/CD pipelines tailored for model, prompt, and agent releases, including automated evaluation and regression gates.
• Implement observability, tracing, drift detection, and cost attribution for LLM and agent workloads.
• Own reliability, scalability, and incident response for daily customer‑facing agentic workflows.
• Create and maintain evaluation pipelines that run on every commit, ensuring only green builds are released.
• Optimize inference cost, latency, and throughput across cloud providers and self‑hosted models.
• Harden the stack for security, secrets management, data residency, and auditability.
• Shape the developer experience to enable AI engineers to ship safely and quickly.

Required profile

• Strong DevOps/SRE background with production‑grade reliability and incident‑response experience.
• Hands‑on exposure to LLMs, Retrieval‑Augmented Generation pipelines, vector databases, or agentic systems.
• Solid understanding of security, networking, and data‑protection fundamentals.
• Engineering mindset focused on correctness, automation, and reproducibility.
• Ability to bridge infrastructure, AI, and product concerns.
• Clear written communication for runbooks, post‑mortems, and dashboards.

Required skills

• Google Cloud Platform (GCP)
• Amazon Web Services (AWS) or Microsoft Azure
• Kubernetes
• Terraform
• CI/CD pipeline tooling
• Large Language Models (LLMs)
• Retrieval‑Augmented Generation (RAG) pipelines
• Vector databases
• Security, networking, and data‑protection concepts