MLOps Architect for AI Production Platforms

Dailymotion; Paris; À distance; EUR 90 000 - 120 000; Plein temps

Résumé du poste

Dailymotion is hiring its first dedicated Devops / MLOps Engineer to own the ML infrastructure in production. You will connect ML engineers with robust tooling and scalable GPU resources in an AI-first environment where fast, reliable model shipping is core to the business.

You will drive ML CI/CD, monitor prod models, and act as the technical mediator between ML and Backbone teams, ensuring a calm, outcome‑oriented approach while reducing technical debt and enabling large-scale experimentation.

Qualifications

• Production ML/DevOps experience essential.
• GCP: Vertex AI, GKE, GCS, BigQuery expertise required.
• Fluent in French and English.
• Experience with GitOps and Kubernetes in production.

Responsabilités

• Empower ML engineers with tooling and infra to iterate autonomously.
• Accelerate time-to-market for production ML products.
• Own ML CI/CD and adapt frameworks for ML needs.
• Monitor and troubleshoot models in production without staging mirrors.
• Enable scalable ML experimentation and A/B testing.
• Deliver ML building blocks (MLflow, Kubeflow, KubeRay) and manage GPU infra.
• Tackle technical debt and set foundations for next steps.
• Bridge ML and Backbone teams to align solutions.
• Handle on-call and post-mortems.

Connaissances

MLOps

DevOps

GCP

Kubernetes

Python

Bash

Go/Rust

GitOps

Terraform

CI/CD

Monitoring

MLflow

Kubeflow

KubeRay

Linux/Containers

Outils

FluxCD

ArgoCD

Docker

Helm

Prometheus

Datadog

Looker

Vertex AI

GKE

GCS

BigQuery

Description du poste

Dailymotion is hiring its first dedicated Devops / MLOps Engineer to own the ML infrastructure in production. You will connect ML engineers with robust tooling and scalable GPU resources in an AI-first environment where fast, reliable model shipping is core to the business.

You will drive ML CI/CD, monitor prod models, and act as the technical mediator between ML and Backbone teams, ensuring a calm, outcome‑oriented approach while reducing technical debt and enabling large-scale experimentation.