MLOps Engineer

ixolabs.ai; France; Sur place; EUR 31 850 - 79 625; Temps partiel

Résumé du poste

ixolabs.ai is seeking an MLOps Engineer for AI training to build, ship, scale, and recover ML pipelines in production. You will work on pipelines, registries, and rollout plans beyond notebook code, contributing to robust automation and reliable deployments.

The role is remote as an independent contractor, with 10-25 hours per week, starting immediately. You will collaborate across cloud platforms and Kubernetes-based serving environments.

Qualifications

• 5 years in MLOps, ML platform engineering, or production ML systems.
• Strong knowledge of Kubernetes-based ML serving and cloud ML platforms.
• Experience with feature stores and model registries.
• Comfort with Python, plus a systems language (Go, Rust, or Java) for platform work.

Responsabilités

• Generate and evaluate instruction-response pairs for ML pipelines, feature stores, and model registries.
• Review AI-generated code using Kubeflow, MLflow, SageMaker, Vertex AI, and Azure ML.
• Provide feedback on CI/CD for ML, model packaging, and reproducibility.
• Validate AI handling of model serving (Triton, TorchServe, BentoML, KServe) and autoscaling.
• Evaluate observability for ML (Evidently, WhyLabs, custom monitoring) and drift detection.
• Identify subtle issues in pipeline DAGs, feature/serving skew, and shadow deployments.

Connaissances

MLOps experience

Kubernetes ML serving

Docker & Helm

Python

GPU scheduling

Outils

Kubeflow

MLflow

Feast

Tecton

SageMaker

Vertex AI

Azure ML

Description du poste

About this role

ML in production lives or dies by the unglamorous work of pipelines, registries, and rollback plans — and AI assistants tend to focus on the model and ignore everything around it. As an MLOps Engineer for AI training, you will help AI generate MLOps code that ships, scales, and recovers, not just the notebook code that demos well.

Key Responsibilities

• Generate and evaluate instruction-response pairs covering ML pipelines, feature stores, and model registries.
• Review AI-generated code using Kubeflow, MLflow, SageMaker, Vertex AI, and Azure ML.
• Provide feedback on CI/CD for ML, model packaging, and reproducibility.
• Validate AI handling of model serving (Triton, TorchServe, BentoML, KServe) and autoscaling.
• Evaluate observability for ML (Evidently, WhyLabs, custom monitoring) and drift detection.
• Identify subtle issues in pipeline DAGs, feature/serving skew, and shadow deployments.

Ideal Qualifications

• 5 years in MLOps, ML platform engineering, or production ML systems.
• Deep familiarity with Kubernetes-based ML serving and at least one major cloud ML platform.
• Strong grasp of Docker, Helm, and infrastructure-as-code for ML workloads.
• Experience with feature stores (Feast, Tecton) and model registries.
• Comfort with Python, plus a systems language (Go, Rust, or Java) for platform work.
• Familiarity with GPU scheduling and distributed training orchestration is a plus.

Project Timeline

• Start Date: Immediate
• Duration: Ongoing
• Commitment: Flexible, 10-25 hours/week

Contract & Payment Terms

• Independent contractor agreement
• Remote work — anywhere in eligible locations
• Weekly payment via Stripe or bank transfer
• Flexible hours