Senior GPU Platform SRE for ML Inference on Kubernetes

Criteo; Grenoble; Hybride; EUR 90 000 - 120 000; Plein temps

Avantages offerts par ce poste

Hybrid work model

Health benefits

Mentorship & career development

Résumé du poste

At Criteo, the Platform Core group builds the foundational infrastructure powering our global advertising platform. We are expanding with a new GPU-focused team to scale ML training and inference workloads using Ray on Kubernetes and NVIDIA Triton Inference Server.

You will join the GPU team as a Site Reliability Engineer, designing, operating, and scaling infrastructure for high-performance model serving and real-time decisioning across services.

Qualifications

• 5+ years of experience in backend engineering, SRE, or platform engineering focusing on distributed systems.
• Strong experience with Kubernetes including workload scheduling and dynamic provisioning.
• Hands-on experience with GPU-based workloads in production for ML training or inference.
• Strong software engineering skills in C#, Python, Go or similar for reliable distributed systems.
• Experience building or operating production-grade infrastructure with performance, scalability and reliability requirements.
• Interest in automation, observability, and scalable systems.

Responsabilités

• Design, operate, and scale infrastructure powering ML training and inference workloads.
• Build and operate scalable Ray clusters on Kubernetes.
• Improve provisioning, observability, reliability, and efficiency of ray-as-a-service environments.
• Operate and optimize large-scale inference platforms using NVIDIA Triton.

Connaissances

Kubernetes

Distributed systems

Backend engineering

C#

Python

Go

GPU workloads

SRE practices

Outils

Ray on Kubernetes

NVIDIA Triton Inference Server

Description du poste

At Criteo, the Platform Core group builds the foundational infrastructure powering our global advertising platform. We are expanding with a new GPU-focused team to scale ML training and inference workloads using Ray on Kubernetes and NVIDIA Triton Inference Server.

You will join the GPU team as a Site Reliability Engineer, designing, operating, and scaling infrastructure for high-performance model serving and real-time decisioning across services.