> Markdown version of [/jobs/ext/4729326-remote-lead-mlops-engineer-databricks](https://www.wearedevelopers.com/jobs/ext/4729326-remote-lead-mlops-engineer-databricks). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # REMOTE - Lead MLOps Engineer (Databricks) - **Company:** State Farm Insurance - **Location:** Bloomington, IL, United States (Remote available) - **Experience:** Expert - **Salary:** $100,000 - $200,000 - **Contract:** Permanent contract - **Skills:** Application Layers, Application Performance Management, Command-Line Interface, Code Review, Continuous Integration, Data Validation, Data Infrastructure, Software Debugging, Python (Programming Language), Open Source Technology, SQL Databases, Strategies of Testing, Apache Spark, Technical Debt, Agentic-AI, Data Lakes, Pyspark, Data Lineage, Deployment Automation, Machine Learning Operations, Drift Detection, Databricks - **Published:** November 15, 2026 - **Apply:** https://www.careerboard.com/us/en/find-jobs-in-United-States/-1C82ADDFEF3113A088/ ## About the Role * Databricks (Non-Negotiable) - At least a year running Databricks in production at the application layer. You've designed and operated medallion architecture, and you know Delta Lake, Unity Catalog, and PySpark cold. * Production ML (Non-Negotiable) - You've kept models healthy in production, not just trained them and handed them off. You've been paged, written the RCA, and caught drift and silent degradation before they became someone else's problem. * Agentic-Native (Non-Negotiable) - GenAI and agentic tools are already part of how you work every day, not an experiment. You're at home on the command line and always looking for the next thing to automate. ## Description * Data Foundations - Design and run medallion architecture (Bronze ? Silver ? Gold) with solid data quality, schema management, and Delta Lake tuning. Everything downstream depends on getting this right. * Application Performance - Hunt down bottlenecks in Spark queries, partitioning, caching, and Databricks cluster sizing, then fix them and keep them from coming back. * Unity Catalog & Lineage - Own schemas, access controls, and project-level tagging, with lineage-driven ticketing so upstream breaks surface right away. * MLOps/MLflow 3.0 - Take models through their full lifecycle, from experiment to production to retirement, including retraining and drift detection. * FinOps - Make Databricks cost visible by default: track, alert on, and optimize spend for every project. * ET Handoff - Define and hold the line on done : low L3 incident rate, complete runbooks, self-healing where possible, and governance and FinOps buttoned up. When a project crosses the handoff plane, it's ready. * Patterns & Practices - Turn what you learn into templates and standards that make every future project better. * Agentic Integration - Bring agentic AI into everything you do: data validation, pipeline generation, testing, monitoring, documentation. If it can be augmented, it should be. * Technical Leadership & Collaboration - Set technical direction, mentor engineers, and pair with pod teams on architecture, code reviews, and debugging. You teach by doing., * MLflow 3.0 - Hands-on with the full lifecycle: registry, experiment tracking, deployment automation, retraining pipelines, and drift detection. * FinOps - You've tracked and optimized Databricks spend at the project level, and you treat compute budgets with the same rigor as SLAs. * Data Lineage & Observability - You've built or run systems that track lineage end to end and open tickets automatically on upstream breaks, schema changes, or SLA misses. * Production Handoff - You've prepared systems for operations teams with runbooks, self-healing patterns, incident targets, and governance checklists. * Systems Thinking - You see the whole system, not just the model or the pipeline. You've read Hidden Technical Debt in Machine Learning Systems, and you've lived it. * Engineering - Production-grade Python, PySpark, SQL; CI/CD for data and ML pipelines; and testing strategies for non-deterministic systems. * Omnigent (Nice to Have) - Bonus points if you've worked with Omnigent, Databricks' open-source agent meta-harness