AI Platform Engineer – Agentic Systems & Cloud

Project overview

For our existing client, we are looking for two experts (m/f/d) with a focus on backend architectures and Agentic AI as soon as possible.

Framework:

• Project start: ASAP
• Duration: until the end of the year
• Option for extension: Yes
• Work location: Remote/Cologne
• Remote share: 60%
• Number of resources: 2

Responsibilities

Our Requirements

• You will conceptualize and implement future-proof multi-agent workflows, agent skills, tool calling via Model Context Protocol (MCP), security guardrails, and human-in-the-loop checkpoints.
• You will build agent skills that independently and reliably translate structured technical requirements into functioning software.
• You will set up secure Azure infrastructures with Terraform (Private Endpoints, Azure Container Apps, Container Registry, Key Vault, PostgreSQL, Monitoring) and develop reusable, versioned Terraform modules with Azure Policy Guardrails.
• You will design automated pipelines in Bitbucket Pipelines with OIDC federation to Azure (without stored secrets), container image scans, and automated test suites.
• You will develop backend services and MCP servers in Java and Python, strictly adhering to input validation and comprehensive audit logging.
• You will operate LLM-based coding agents in isolated containers, connecting modern language models via Azure AI Foundry.
• You will seamlessly integrate Bitbucket, Jira, and Microsoft Teams into automated workflows via their APIs.
• You will keep architectural decisions transparent as Architecture Decision Records (ADRs) and actively share your knowledge within the team.
• At least 6 years of in-depth practical experience in backend or platform engineering.
• Demonstrable experience in designing agentic architectures (multi-agent workflows, agent skills, tool calling, MCP servers, guardrails).
• Extensive knowledge of Microsoft Azure – especially Azure Container Apps (or comparable container platforms), Private Endpoints, Key Vault, and Managed Identities.
• Solid production experience with Terraform (module design, remote state, policy as code).
• Routine in pipeline creation with secretless, OIDC-based cloud access (ideally in Bitbucket Pipelines).
• Strong programming skills in Java and Python, REST API design, and relational data modeling in PostgreSQL.
• Consistent focus on least privilege principles, sandboxing, input validation, and audit trails.
• Practical experience with Azure AI Foundry or Azure OpenAI Service (model deployments, quota management).
• Familiarity with open-source coding agent harnesses and their extension architectures.
• Knowledge of common agent frameworks (e.g., LangGraph, Semantic Kernel, or Spring AI).
• Experience with TypeScript and Node.js.
• End-to-end testing with Playwright within CI/CD pipelines.
• Knowledge of Jira Automation and MS Teams notifications/webhooks.
• Industry experience in regulated environments

Project details

• Start: ab sofort
• Duration: 2 Monate
• Remote: hybrid
• Location: köln, Deutschland, DE
• Languages: German