MLOps Engineer (Relocation to Serbia)

Role Overview

Build and operate an internal AI platform that helps developers ship AI-powered (LLM-based) services faster. The platform covers model connectivity, prompt testing and evaluation, monitoring/observability, and the underlying AI infrastructure.

Responsibilities

• Build and operate AI platform infrastructure for faster delivery of LLM-based services.
• Implement and maintain Kubernetes-based runtime environments (including AKS) for AI workloads.
• Manage infrastructure as code using Terraform (modules, multi-environment setups, and CI/CD automation).
• Support LLM workflows including:
  • RAG and retrieval components
  • Agents
  • Prompt experimentation and evaluations
  • Deployment patterns
• Integrate and operate tooling such as:
  • Azure AI Foundry
  • LiteLLM
  • Langfuse
  • MLflow
• Orchestrate ML/LLM pipelines using Kubeflow Pipelines and/or Argo Workflows (build, deploy, evaluate).
• Improve platform reliability and observability:
  • monitoring, logging, tracing
  • cost/performance signals
• Collaborate with developers to streamline developer experience (DX): APIs, templates, documentation, “golden paths”, and automation.

Requirements

• Hands-on Kubernetes production experience (preferably with AKS).
• Strong Terraform expertise (IaC best practices, multi-environment setups).
• Practical experience supporting ML/LLM workloads in a platform or DevOps/MLOps context.
• Proficiency in Python for automation, scripting, and supporting APIs/evaluation tooling.
• Understanding of CI/CD, release processes, and production-grade operations.
• Ability to deliver pragmatically under tight timelines.

Nice to Have

• Experience building internal developer platforms / paved roads.
• Familiarity with LLM evaluation frameworks, prompt testing workflows, and LLM observability.
• Exposure to RAG architectures, vector databases, and agentic patterns.
• Experience with Kubeflow, Argo, and ML lifecycle tooling.

Engagement & Team

• Long-term B2B contract.
• Team of 5; 3 additional AI Platform Engineers being added.

Location / Timezone

• Remote work from Croatia, Poland, Portugal, or Serbia.
• European working hours; meetings occasionally up to 10:00 AM PST (US overlap).