Remote Senior MLOps Engineer — Production-Scale AI

GCS; United Arab Emirates; On-site; AED 551,000 - 845,000; Full time

Job summary

GCS is hiring a Senior MLOps Engineer to own the infrastructure that trains, serves, and scales large language models in production. You will manage self-hosted models, guardrails, and a reliable, cost-efficient AI platform from a remote setup.

Lead end-to-end model lifecycles, implement robust monitoring, and drive continuous improvement across Blended cloud and on-prem environments. The role emphasizes on-call incident handling, distributed training, and scalable ML pipelines using MLflow,

Qualifications

• 5+ years in MLOps, ML infra, or ML engineering with end-to-end production lifecycle.
• Hands-on experience serving large models at scale (vLLM, Triton, TGI) and GPU optimization.
• Strong Kubernetes, Docker, and IaC (Terraform).
• On-call experience with ML services and incident handling.

Responsibilities

• Deploy and scale self-hosted open-weight models using vLLM, Triton, or TGI to hit latency and cost targets.
• Operate a multi-provider gateway with token accounting and cost-aware routing.
• Build pipelines for fine-tuning, evaluation, versioning, and CD (MLflow, SageMaker, Kubeflow), incl distributed training (DeepSpeed, FSDP, Accelerate).
• Own production reliability: monitoring, logging, alerting, incident response, safe rollback.

Skills

MLOps

Model lifecycle

Kubernetes

Docker

Terraform

Quantization

vLLM

Triton

TGI

SageMaker

Kubeflow

MLflow

DeepSpeed

Accelerate

FSDP

Tools

MLflow

Kubeflow

SageMaker

DeepSpeed

Accelerate

Triton

vLLM

Job description

GCS is hiring a Senior MLOps Engineer to own the infrastructure that trains, serves, and scales large language models in production. You will manage self-hosted models, guardrails, and a reliable, cost-efficient AI platform from a remote setup.

Lead end-to-end model lifecycles, implement robust monitoring, and drive continuous improvement across Blended cloud and on-prem environments. The role emphasizes on-call incident handling, distributed training, and scalable ML pipelines using MLflow,