ML Ops

Responsibilities: ML Pipeline Design & Automation Build and maintain CI/CD & CT (Continuous Training) pipelines for ML models using Azure DevOps and Databricks Asset Bundles. Automate data preprocessing, training, inference and retraining workflows for large-scale ML deployments. Implement incremental backfills and rolling window retraining for time-series forecasting. Deployment & Infrastructure Design job clusters and compute policies in Databricks for optimal cost-performance trade-offs. Implement multi-environment deployment flows (Dev → QA (stage) → Prod) with approvals and rollback strategies. Deploy ML models to production with monitoring hooks for performance and drift detection. Data & Model Governance Integrate with Unity Catalog for secure, compliant data and model storage. Set up model versioning, lineage tracking and reproducibility using MLflow. Establish dataset and feature versioning using tools like Databricks Feature Store. Monitoring & Observability Implement structured logging for model metrics, system performance and data quality checks. Integrate monitoring tools (e.g., Azure Application Insights) for alerting and dashboards. Develop automated retraining triggers based on performance degradation. Qualifications: B.Tech/BCA or any Graduation Experience with time-series forecasting at scale (e.g., Prophet, Sarima, XGBoost). Experience in retail demand forecasting and/or energy sector analytics. Knowledge of feature engineering at scale with distributed systems.