MLOps Engineer

Overview

Senior MLOps / ML Engineer role to own, develop, and maintain ML deployment and data processing workflows, spanning an Argo-based workflow orchestration platform and Python-based frameworks for large-scale AI model inference, within the CARIAD COE in Bangalore.

What you'll do

• Design and maintain Argo Workflows / Argo Events–based ML and data processing pipelines across dev / QA / staging / prod environments.
• Design and maintain WorkflowTemplates, DAGs, fan-out patterns, scheduled workflows, and event-driven triggers following GitOps best practices.
• Design and develop Python-based workflows and components for ML data processing and model inference.
• Work with Databricks and Spark-based workloads for data processing, transformation, and pipeline execution.
• Build and maintain Docker images for ML and GPU-accelerated workloads, including multi-stage builds and container optimization.
• Develop scalable inference workflows using Ray where required.
• Integrate with Azure ML and Azure cloud services for model versioning, model artifacts, training workflows, and deployment.
• Maintain CI/CD pipelines with automated testing, quality gates, and environment-specific deployment processes.
• Automate deployment, monitoring, and rollback of machine learning and deep learning models on production Kubernetes environments.
• Manage model versioning and artifact movement from cloud ML workspaces and storage systems to production workloads.
• Maintain configuration systems and reusable workflow components to ensure reproducible pipeline execution across environments.
• Monitor model and pipeline performance, data processing health, resource utilization, and infrastructure reliability.
• Troubleshoot production issues involving failed workflows, data bottlenecks, resource contention, GPU utilization, and latency regressions.
• Identify reliability and performance issues proactively and implement improvements to ML and data processing workflows.
• Collaborate with data scientists and engineering teams to improve model deployment, inference, and pipeline reliability.

What you'll need

• Python: strong programming and problem-solving skills; experience with application development, scripting, and Python packaging.
• Workflow design: ability to design reliable workflows with dependencies, parallel execution, retries, failure handling, scheduling, and event-driven execution.
• Argo Workflows / Argo Events: understanding of WorkflowTemplates, DAGs, event-driven workflows, and workflow execution.
• Databricks / data processing: hands-on experience with Databricks, Spark/Spark SQL, or similar large-scale data processing platforms.
• Kubernetes: understanding of workload orchestration, resource management, namespaces, and production troubleshooting.
• Git / version control: strong understanding of Git or equivalent, including branching, merging, pull requests, and maintaining code/configuration changes.
• Docker: containerization, multi-stage builds, and optimization of ML workloads.
• CI/CD: experience designing pipelines with automated testing, quality gates, artifact management, and deployment processes.
• Azure cloud: working knowledge of Azure ML, Azure Blob Storage, AKS, and Azure-based ML workflows.
• Problem solving: ability to troubleshoot unfamiliar technical problems and identify practical solutions.

Nice to have

• Ray: distributed data processing and inference at scale.
• Deep learning frameworks: PyTorch, TensorFlow; model formats such as ONNX, SavedModel, and TorchScript.
• Helm / ArgoCD and GitOps-based deployment practices.
• Prometheus / Grafana and production observability.
• Experience with Azure Pipelines or similar CI/CD platforms.
• Experience with ML model monitoring, data drift, and model performance monitoring.
• Large-scale data processing: experience working with large volumes of structured, unstructured, image, or video data, particularly for ML/AI workloads.

Details

• Location: Bangalore, India.