Data Infrastructure Engineer for Large-Scale AI Platform (Hybrid)

Mistral.ai; Paris; Hybrid; EUR 90,000 - 150,000; Full time

Benefits offered by this job

Healthcare coverage

Parental leave

Relocation support

Retirement plans

Wellness programs

Meal and transport allowances

Location-specific perks

Job summary

Mistral AI is building the next generation of data infrastructure for large-scale model training and fine-tuning. You will own core components for compute and storage, enabling scalable, secure access for MLOps and research across multi-cluster environments.

The role emphasizes migrating away from legacy schedulers to production-grade pipelines and participating in on-call rotations for critical training jobs. Hybrid/remote options with EU hubs are available.

Qualifications

• 4+ years of experience in Data Infrastructure, MLOps, or Infrastructure Engineering.
• Experience with large-scale compute and storage platforms is preferred.
• Proficiency in Python and Kubernetes-native tooling.

Responsibilities

• Build and scale massive distributed compute and storage systems.
• Architect multi-cluster orchestration layers to optimize workload placement.
• Design storage formats for exabyte-scale datasets.
• Contribute to internal training platform across Kubernetes and SLURM environments.
• Implement data lineage and metadata visibility as pipelines grow.
• Operate cloud-native deployments with reliable, scalable workflows.

Skills

Data infrastructure

MLOps

Distributed systems

Infra engineering

Tools

Kubernetes

SLURM

Python

Job description

Mistral AI is building the next generation of data infrastructure for large-scale model training and fine-tuning. You will own core components for compute and storage, enabling scalable, secure access for MLOps and research across multi-cluster environments.

The role emphasizes migrating away from legacy schedulers to production-grade pipelines and participating in on-call rotations for critical training jobs. Hybrid/remote options with EU hubs are available.