Databricks launches Genie Code, an autonomous AI agent for data engineering and analytics. The agent can perform complex tasks such as building data pipelines and debugging code.
Databricks has introduced Genie Code. This autonomous AI agent transforms the way organizations work with data. Genie Code can perform complex tasks such as building data pipelines, debugging code, delivering dashboards, and managing production systems. During field tests, Genie Code achieved a success rate more than twice as high as other coding agents, Databricks claims.
Genie Code complements Genie, a tool that allows knowledge workers to talk to their data and get reliable answers instantly. Just like Genie, the coding agent takes on the engineering work to bring new ideas to production faster.
Experienced ML engineer
According to Databricks, Genie Code performs at the level of an experienced machine learning engineer. The agent supports full ML workflows. Genie Code reasons through complex problems to plan, build, and deploy models. Experiments are logged in MLflow, and serving endpoints are tuned for maximum performance.
The agent accounts for the differences between staging and production environments, builds workflows for change data capture, and applies data quality rules. In the background, it monitors Lakeflow pipelines and AI models to assess failures and investigate anomalies. The agent independently analyzes agent traces to fix hallucinations and tunes resource deployment before a human needs to intervene.
Acquisition of Quotient AI
Databricks also announces that it has acquired Quotient AI. Quotient automatically monitors agent performance. It measures the quality of answers, identifies regressions at an early stage, and locates points of failure. These insights then feed into a reinforcement learning cycle that ensures agents continuously improve.
The founders of Quotient bring expertise in evaluating AI coding systems. They previously led the quality improvement of GitHub Copilot. By building these capabilities into Genie Code, Databricks ensures that data and AI systems continue to evolve once they are running in production.