Databricks MLops Engineer

Role: Databricks MLOps Engineer

Experience: 6+ Years of IT Experience (3+ years building ML pipelines on Databricks)

Location: All EXL Locations (Hybrid)

We are seeking a Databricks MLOps Engineer with 6+ years in ML engineering, MLOps, data engineering, including 3+ years building ML pipelines on Databricks, to take models from notebooks to tested, monitored production. The role covers feature pipelines, training automation, model registry, serving, monitoring and CI/CD for ML, and supports LLM and GenAI workloads with clear cost controls.

• Build feature pipelines on Delta tables and manage features in Unity Catalog feature tables.
• Automate training and retraining with Databricks Workflows, including parameterized and scheduled runs.
• Track experiments and register, version and promote models with MLflow and the Unity Catalog model registry, using aliases for champion and challenger models.
• Deploy models to Databricks Model Serving (real-time) and batch inference jobs; run A/B and shadow deployments.
• Monitor models with Lakehouse Monitoring and inference tables: data drift, prediction drift, latency, error rates and alert thresholds.
• Implement CI/CD for ML with Databricks Asset Bundles and GitHub Actions or Azure DevOps: tests, staged promotion and rollback.
• Support LLM and GenAI workloads: serving endpoints, AI Gateway rate limits, model-tier selection and token usage tracking per endpoint.
• Partner with data scientists to turn notebooks into tested, packaged production code.

Required Technical Skills Core Technologies

• Databricks Lakehouse Platform
• MLflow (tracking, registry, deployment)
• Python for production
• PySpark and Delta Lake

ML Engineering

• Feature engineering and Unity Catalog feature tables
• Training automation with Databricks Workflows
• Model Serving and batch inference
• Champion/challenger promotion, A/B and shadow testing

Monitoring & Governance

• Lakehouse Monitoring and inference tables
• Data and prediction drift detection, alerting and retraining triggers
• Unity Catalog model governance and access control

DevOps

• Databricks Asset Bundles
• CI/CD with GitHub Actions or Azure DevOps
• Unit and integration testing (pytest)
• Git-based workflows

LLM / GenAI Operations

• Serving endpoints and AI Gateway
• Token usage tracking, model-tier selection and cost per call
• Vector Search for RAG use cases

Required Qualifications

• Bachelor's or Master's degree in Computer Science, Data Science, Engineering or a related field.
• 6+ years of experience in ML engineering, MLOps or data engineering with ML delivery.
• 3+ years of hands-on experience building ML pipelines on Databricks (Workflows, MLflow, Model Serving or batch inference).
• Multiple models deployed to production with monitoring, retraining and rollback in place.
• Strong Python engineering practices: packaging, testing, logging and error handling.

Preferred Qualifications

• Databricks Certified Machine Learning Associate or Professional.
• Experience with Feature Store and online tables.
• Experience supporting LLM or GenAI applications in production.
• Docker, REST APIs and basic cloud networking for serving endpoints.

Soft Skills

• Ability to explain model behaviour, drift and risk to business stakeholders in plain language.
• Strong collaboration with data scientists, data engineers and product teams.
• Experience working in Agile/Scrum environments.
• Clear documentation of pipelines, runbooks and release processes.

Nice to Have

• Azure Machine Learning or Amazon SageMaker
• Kubernetes
• Media, telecom or sports analytics use cases (churn, recommendations, audience)