Databricks introduces an autonomous AI agent Genie Code that automates data engineering tasks to manage complex data workflows.
Databricks, the Data and AI Company, which has over 20,000 enterprises globally has introduced Genie Code, an autonomous AI agent that revolutionizes data labor. Complex operations like creating pipelines, troubleshooting errors, shipping dashboards, and maintaining production systems can all be completed using Genie Code.
Due to their lack of access to crucial context, such as lineage, usage patterns, and business semantics, current agentic coding tools struggle to complete data tasks. In order to guarantee the high standards of correctness and governance necessary for production environments, Genie Code assists teams in bridging the context gap.
According to Ali Ghodsi, Co-founder and CEO of Databricks, in the past six months, software development has shifted from code-assistance to full agentic engineering. He said that Genie Code brings this transformation to data teams and that they are going from a world where data professionals are assisted by AI to one where AI agents execute the work, under human guidance. And they call this Agentic Data Work. “It will fundamentally change how enterprises make decisions,” he said.
Leading coding agents' success rate on real-world data jobs was doubled by Genie Code. 80% of new datasets on Databricks are already built by AI rather than humans. Further, Genie Code is considered as a recent addition to Genie, enabling any knowledge worker to communicate with their data and receive reliable responses quickly by utilizing the context and semantics recorded by Unity Catalog. This method is extended to data professionals by Genie Code, which manages the intricate engineering needed to move all company data from concept to production.
Additionally, Quotient AI, a pioneer in evaluation and reinforcement learning for AI agents, is acquired by Databricks today in order to integrate continuous evaluation directly into Genie and Genie Code. It is a reinforcement learning loop that maintains agents getting better over time is fed by Quotient's automatic monitoring of agent performance, which measures answer quality, detects regressions early, and identifies failures.
