Agentic AI & MLOps Engineer- LEAD

Role & responsibilities -

• MLOps Platform Ownership: Design, build, and maintain the MLOps infrastructure underpinning the Agentic AI platform model serving, experiment tracking, evaluation pipelines, and deployment automation using Databricks MLflow.

• Model Serving & Lifecycle Management: Manage ML model endpoints on Databricks Model Serving (REST); oversee model versioning, promotion, and deprecation workflows across dev, staging, and production environments.

• LLM Evaluation Pipelines: Build and operate automated evaluation pipelines for LLM outputs accuracy, relevance, groundedness, and safety metrics integrated with MLflow experiment tracking.

• CI/CD for AI Workloads: Own and evolve the GitHub-based CI/CD pipeline for agent and model code — automated testing, quality gates, environment promotion, and deployment to production.

• Observability & Monitoring: Implement and maintain end-to-end observability for agentic workflows using Application Insights and Azure Monitor — execution traces, token usage, latency, error rates, and alerting.

• Data Pipeline Automation: Automate data ingestion pipelines into Databricks Unity Catalog and Azure AI Search (RAG document ingestion); ensure data quality, freshness, and governed access.

• Secrets & Identity Management: Maintain secure credential management via Azure Key Vault and Managed Identity; enforce zero-secrets-in-code policies across all platform components.

• Agentic Engineering Contribution: Contribute to agentic workflow development in LangGraph — particularly around tool-call reliability, error handling, retry logic, and workflow resilience.

• Governance & Audit: Maintain the LLM output review store (Cosmos DB / Blob Storage) for governance sign-off; support audit and compliance requirements for AI outputs.

Required candidate profile-

Strong in coding, Azure AI Stack, Langraph, LLM

• MLOps & Model Lifecycle: 3+ years of MLOps experience; deep expertise in Databricks MLflow for experiment tracking, model registry, and evaluation.

• Databricks Platform: Strong hands-on experience with Databricks — Unity Catalog, SQL endpoints, Model Serving (REST), and Delta Lake.

• CI/CD for AI/ML: Proven experience building and maintaining GitHub Actions or equivalent CI/CD pipelines for ML and AI workloads — including automated testing, staging, and production promotion.