Architect - DevOps 4D

Overview

Own the CI/CD delivery pipeline, deployment automation, and observability and guardrail layer for ML and GenAI/LLM applications on Azure and Databricks. Infrastructure is provisioned and managed by a separate platform team.

What you'll do

• Build and maintain CI/CD pipelines in GitHub Actions for Python applications, ML models, and GenAI/LLM services.
• Implement branching, versioning, and artifact promotion across DEV, SIT, and PROD environments.
• Automate deployment of Databricks jobs, notebooks, and workflows using Databricks Asset Bundles and the Databricks CLI.
• Automate model deployment to Databricks Model Serving and Azure ML online endpoints, including canary rollout and rollback.
• Integrate MLflow into pipelines for experiment tracking, model registry in Unity Catalog, and stage promotion.
• Add automated quality gates to pipelines: unit tests, linting, container scanning, model validation, and LLM evaluation checks.
• Implement LLM evaluation harnesses in CI, including golden datasets, RAGAS-style retrieval metrics, LLM-as-judge, and regression tests that block bad prompt or model changes.
• Configure guardrails for GenAI applications: input/output filtering, PII redaction, prompt-injection and jailbreak checks, and Azure AI Content Safety.
• Set up observability for models and LLM applications using MLflow Tracing, OpenTelemetry, and Azure Monitor / Application Insights, covering latency, token usage, cost, tool calls, and failure modes.
• Configure monitoring and alerting for data drift, model performance degradation, and endpoint health.
• Manage secrets, service principals, and OIDC-based authentication between GitHub Actions and Azure/Databricks.
• Containerise applications with Docker and deploy to existing AKS or Azure Container Apps environments.
• Work with data scientists and GenAI engineers to move use cases from notebook to production reliably.

What you'll need

• 6–10 years in DevOps, platform, or ML engineering roles.
• Strong hands-on experience building CI/CD with GitHub Actions, including reusable workflows, matrix builds, environments, approvals, and self-hosted runners.
• Solid Python skills, plus Bash, YAML, and Git.
• Hands-on experience with Databricks: Workflows, Unity Catalog, MLflow, Model Serving, Asset Bundles, and Databricks CLI.
• Working knowledge of Azure: Azure ML, ADLS Gen2, Key Vault, ACR, AKS or Container Apps, Entra ID service principals, and Azure Monitor.
• Practical MLOps experience: model packaging, registry, versioning, automated retraining, and drift and performance monitoring.
• Practical LLMOps experience with production GenAI applications, including RAG and/or agent frameworks such as LangChain or LangGraph.
• Hands-on experience implementing LLM guardrails and evaluation, not theoretical familiarity.
• Experience with Docker and container-based deployment.
• Experience with observability tooling and instrumenting applications for tracing and metrics.
• Bachelor's in Computer Engineering, Computer Science, or Information Technology, or Master's in Computer Science.
• Language: English, Advanced (C1), required.

Nice to have

• Azure or Databricks certification.
• Experience in a regulated industry with model governance or audit requirements.
• Exposure to Langfuse or LangSmith, and to inference cost optimisation.
• Familiarity with Azure DevOps pipelines alongside GitHub Actions.

Details

• Location: Prestige Technology Park IV, Bangalore, India.
• Work mode: Hybrid.
• Work shift: Flex Time (India).
• Job type: Regular.