AI Engineer: On-Prem LLMs & Secure Production

Twoday Denmark; København; On-site; DKK 900,000 - 1,300,000; Full time

Job summary

Twoday Denmark is seeking an AI Engineer to advance on-premise LLM solutions for a Danish public-sector project. You will work across the development and operations of LLM-based services, focusing on secure, production-ready deployments and AI agents.

You'll design, implement and test RAG flows, tool-calling and structured outputs, with strong Python, Docker, Linux and CI/CD skills, and collaborate with stakeholders onsite in Aarhus.

Qualifications

• Hands-on experience building production-ready LLM solutions.
• Experience with RAG, prompt design, tool calling and structured outputs.
• Knowledge of local or self-hosted models and inference infrastructure.
• Experience with Docker, Linux, APIs and CI/CD.
• Fluent Danish and English, both spoken and written.
• Ability to work full time onsite in Aarhus.

Responsibilities

• Developing and integrating LLM-based applications and agents.
• Deploying and operating LLM models in on-premise environments.
• Building RAG solutions, structured workflows and validation layers.
• Improving model reliability through guardrails, testing and monitoring.
• Integrating AI services with existing applications and APIs.
• Working with sensitive data and high requirements for security and traceability.

Skills

Python

LLM frameworks

RAG & prompts

Structured outputs

Monitoring & validation

Hallucination reduction

Danish language

English language

Tools

Docker

Linux

APIs

CI/CD

On-premise deployment

Job description

Twoday Denmark is seeking an AI Engineer to advance on-premise LLM solutions for a Danish public-sector project. You will work across the development and operations of LLM-based services, focusing on secure, production-ready deployments and AI agents.

You'll design, implement and test RAG flows, tool-calling and structured outputs, with strong Python, Docker, Linux and CI/CD skills, and collaborate with stakeholders onsite in Aarhus.