ML DevOps Engineer

You have the unique opportunity to join our team, building an AI/ML platform to develop and train state-of-the-art AI-based cybersecurity agents. This is a hands-on infrastructure work in a security-sensitive environment, in collaboration with real external R&D teams building solutions on this very platform.

We are looking for an experienced engineer to help operate and enhance what has been done already to foster the platform growing even further. You will have the opportunity to collaborate with other internal project teams, and contribute in various ways to other project goals while ensuring scalability, compliance, and excellence in infrastructure services.

Key responsibilities

Design, build, and operate AI/ML infrastructure: GPU cluster management (Incus, LXD fork), Kubernetes (Talos Linux), model registry, inference serving, and data pipelines

Build reproducible, auditable environments with CI/CD and configuration management (Ansible, GitOps)

Provide the platform as an internal and external service: user on-boarding, support, documentation and capacity planning

Deploy, benchmark, and tune open-weight LLMs and other models on our own hardware

Work with R&D technical leads and project managers to keep infrastructure aligned with project content, budget, and schedule

Apply security-by-design in a classified environment: hardening, isolation, patching, compliance

Mentor junior team members and, as the team grows, take on team lead responsibilities, if necessary/possible

This role contributes to the Research and Development project AIDA, building into EU defence capabilities

Ideal candidate must have

3+ years of DevOps or infrastructure engineering experience in production environments

Solid fundamentals in Linux administration, Git, CI/CD, configuration management (e.g. Ansible)

Hands-on experience with containers and Kubernetes in production

Scripting and automation in Python and Bash

Experience operating services for other teams like monitoring, incident handling, documentation

Strong communication skills, both spoken and written, including proficiency in English

Willingness to apply for security clearance

Readiness to work on-site in Tallinn, with the option for partial hybrid work if desired

Ideal candidate would have experience in or would be willing to learn

ML workloads: GPU drivers/CUDA, model serving (vLLM, Ollama, llama.cpp), experiment tracking (Mlflow)

Infrastructure as code, GitOps, and observability tooling (Terraform/OpenTofu, Argo CD/Flux, KubeFlow, Prometheus/Grafana)

Cybersecurity practices, methodologies and standards

Working in restricted network environments

You don't need to tick every box, but if you're a strong DevOps engineer who is curious about ML/AI and wants to see your work used in real cybersecurity domain then we are looking for you! 