Senior MLOps Engineer (OpenShift / AI Platform)

Company Description

OBSS is Turkey’s largest technology consultancy, pioneering product development and technology consulting for an AI-native future. With more than 20 years of experience and offices in London, Istanbul and Ankara, we empower leading companies in banking and finance, insurance, e-commerce, retail, and the start-up ecosystem. Our team of over 1000 talented professionals thrives on collaboration, innovation, and continuous learning.

About the Role

We are looking for a highly skilled Senior MLOps Engineer to architect, scale, and maintain our enterprise-grade machine learning infrastructure. In this role, you will bridge the gap between data science and core infrastructure. You will build reliable pipelines across a cutting-edge hybrid ecosystem spanning on-premises OpenShift multi-GPU environments and major global cloud providers.

The ideal candidate understands how to extract maximum performance from NVIDIA GPU clusters bare-metal/on-prem, while simultaneously deploying scalable workflows across AWS, Azure, GCP, and Huawei Cloud.

Key Responsibilities

• Hybrid Infrastructure & Orchestration: Manage and scale production ML workloads across containerized systems, specializing in Red Hat OpenShift for on-premises deployments and managed Kubernetes solutions (EKS, AKS, GKE, CCE) in the cloud.
• GPU Compute Optimization: Maximize the performance of physical on-premises NVIDIA multi-GPU clusters. Configure, monitor, and partition resources using NVIDIA CUDA, TensorRT, and MIG (Multi-Instance GPU) profiles.
• Multi-Cloud Scalability: Deploy, monitor, and manage models across a hybrid cloud topology utilizing native AI services including AWS SageMaker, Azure ML, GCP Vertex AI, and Huawei Cloud (ModelArts).
• CI/CD/CT Pipeline Engineering: Design and build fully automated end-to-end Machine Learning pipelines—covering Continuous Integration, Continuous Deployment, and Continuous Training (CI/CD/CT)—using MLflow, Kubeflow, Argo Workflows, or GitLab CI.
• Model Serving & Low-Latency APIs: Establish robust runtime environments for high throughput batch and real-time model inference using Triton Inference Server, vLLM, KServe, or FastAPI.
• Observability & Governance: Implement strict logging, tracing, and metric frameworks to monitor system performance, resource optimization, data drift, and model decay using Prometheus, Grafana, and OpenTelemetry.

Technical Requirements

• 3–6 years of professional experience in MLOps, ML Platform Engineering, or DevOps supporting heavy-compute production workloads.
• Hands-on experience managing enterprise Kubernetes platforms, specifically Red Hat OpenShift, along with low-level NVIDIA hardware/software stack alignment (CUDA, drivers, drivers operator).
• Proven tracking deploying architectures across AWS, Azure, or GCP, combined with familiarity or hands-on experience in Huawei Cloud infrastructure (Ascend NPU integration is a massive plus).
• Proficient with Terraform, Ansible, or Pulumi for immutable infrastructure delivery.
• Expert-level Python and robust Linux shell scripting capabilities.
• Familiarity with feature stores, data versioning tools (DVC), or vector databases for Generative AI deployment setups.

Nice-to-Have Qualifications

• Experience with Huawei Ascend NPU hardware architecture.
• Deep engineering knowledge of Large Language Model operations (LLMOps), including distributed fine-tuning framework architectures (DeepSpeed, Megatron-LM).
• Red Hat Certified Specialist in OpenShift Administration or recognized Multi-Cloud Professional Architect certifications.

OBSS is proud to be an equal opportunity workplace and is an affirmative action employer. We make recruiting decisions without regard to race, color, religion, national or ethnic origin, age, gender, sexual orientation, marital status, veteran status or disability status.