1. Basic Details

• Start: April

• Duration: 3 months with possibility of prolongment 

• Location: Stockholm

• Workload: 100%

2. Project

• Scope: Build and operationalize Model as a Service on the client AI Platform.

• Collaboration: Head of MLE, MLE Tech Lead, MLE team, AI Platform team.

3. Profile Requirements

Must have‑:

• 5–8+ years across ML engineering / MLOps / platform engineering, with 2+ years specifically on model serving or inference platforms.

• Strong Python and one additional language (Go/Java/TypeScript helpful for gateway/tooling).

• Production experience with at least one of: KServe, Seldon, Triton, BentoML, Ray Serve, vLLM/TGI.

• CI/CD + IaC fluency: GitLab CI/GitHub Actions/Azure DevOps, Terraform, Helm/Kustomize.

• Security mindset: secrets management, network policies, mTLS, image signing, SBOM, vulnerability management.

• Comfortable in regulated environments with audits, approvals, and clear change control

Nice to‑ ‑have:

• Performance tuning for LLM inference 

• Feature stores, vector databases, online/offline evaluation, human feedback pipelines.

• Cost optimization for GPU fleets and autoscaling strategies.

Tooling (can include do not have to): 

• Serving: KServe, Seldon, BentoML, Triton, ONNX Runtime.

• API & Auth: Kong, OAuth2/OIDC, Vault/KMS/HSM.

• CI/CD & IaC: GitLab CI/GitHub Actions, Terraform, Helm/Kustomize.

• Registry: MLflow, model cards, artifact registries.

• Safety: toxicity/PII classifiers, prompt sanitizers, output filters.

• NVIDIA ecosystem experience.
  Experience: 5-8+ years
  Languages: Swedish & English

4. Responsibilities

• Deliver Model as‑ a‑ ‑Service components according to architectural and operational standards.

• Work closely with MLE tech leadership and platform teams to ensure successful implementation.

• Deliverables evaluated by Head of MLE and MLE Tech Lead.