Surprisingly Useful AI Article Enhancements
Databricks has launched Genie Code, an autonomous AI agent designed to handle complex data engineering, data science and analytics tasks from idea through to production.
The company claims the tool more than doubled the success rate of leading coding agents on real-world data science tasks, lifting completion rates from 32.1 per cent to 77.1 per cent.
Genie Code arrives alongside the acquisition of Quotient AI, a company specialising in evaluation and reinforcement learning for AI agents, which Databricks plans to embed directly into its Genie product line.
From Code Assistance To Agentic Data Work
Databricks is positioning Genie Code as a shift in how data professionals interact with AI tooling. Where existing tools function primarily as helpers – writing code, running local tests and iterating – Genie Code is designed to reason through problems, plan multi-step approaches, write and validate production-grade code and maintain the output over time.
The company draws a parallel with the evolution of software engineering tools, which have moved from autocomplete-style assistance to agent-driven development over the past year.
“Software development has shifted from code-assistance to full agentic engineering in the past six months,” Ali Ghodsi, Co-Founder and CEO of Databricks, explained. “Genie Code brings this revolution to data teams. We’re moving from a world where data professionals are assisted by AI to one where AI agents do the work, guided by humans.”
Ghodsi described the approach as “Agentic Data Work”, arguing it will change how enterprises make decisions.
Bridging The Context Gap
According to Databricks, a key limitation of existing agentic coding tools when applied to data tasks is their lack of access to critical context such as lineage, usage patterns and business semantics.
Genie Code is integrated with Unity Catalog, Databricks’ governance layer, which gives the agent access to existing governance policies, access controls, business semantics and audit requirements.
The tool is designed to handle full machine learning workflows end-to-end, from planning and writing models through to deploying them, while logging experiments to MLflow and fine-tuning serving endpoints.
On the data engineering side, Genie Code is built to account for differences between staging and production environments, build workflows for change data capture and apply data quality expectations – tasks that Databricks argues distinguish a senior data architect from a novice engineer.
Genie Code also monitors Lakeflow pipelines and AI models in the background, triaging failures, investigating anomalies and tuning resource allocation without waiting for human intervention.
The agent improves over time through persistent memory, automatically updating internal instructions based on past interactions and coding preferences.
Early Adopters Describe A Shift In Workflow
Several enterprise customers are already using Genie Code in their operations.
Bernie Graham, VP of Data Engineering at SiriusXM, described the tool as a development partner for data teams.
Related: Best Business Laptops for work & school
Related: Best Gaming Laptops
Related: Best Portable Laptop
“At SiriusXM, Genie Code supports everything from authoring notebooks and complex SQL to reasoning through table relationships and debugging pipelines,” Graham remarked. “It acts as a hands-on development partner that helps our data teams deliver high-quality work in less time.”
Emilio Martin Gallardo, Principal Data Scientist for Data Management and Analytics at Repsol, pointed to the tool’s ability to handle complex workflows that would otherwise require manual orchestration.
“Instead of stitching together notebooks, pipelines and models manually, we can hand off complex workflows to an AI partner that understands our data, governance, business context and internal libraries,” Gallardo observed. “It accelerates everything from time series forecasting to production deployment, without sacrificing rigor or control.”
Quotient AI Acquisition Adds Continuous Evaluation
Alongside the Genie Code launch, Databricks has acquired Quotient AI to strengthen the evaluation and monitoring capabilities of its agent products.
Quotient automatically monitors agent performance by measuring answer quality, catching regressions early and identifying failures. It feeds a reinforcement learning loop intended to keep agents improving over time.
The Quotient team brings experience in evaluating AI coding systems, having previously led quality improvement efforts for GitHub Copilot.
By embedding Quotient’s capabilities into Genie Code, Databricks aims to ensure its data and AI systems do not just run in production but continuously improve their performance.
Part Of A Broader Platform Push
Genie Code joins an existing product called Genie, which lets knowledge workers chat with their data and receive answers using the context and semantics captured by Unity Catalog.
Where the original Genie product targets general business users, Genie Code is aimed at data professionals who need to build and maintain production systems.
Databricks, which counts more than 20,000 organisations worldwide as customers including Adidas, AT&T, Bayer, Block, Mastercard and Rivian, offers a unified Data Intelligence Platform that includes Lakebase, Genie, Agent Bricks, Lakeflow, Lakehouse and Unity Catalog.
The company is headquartered in San Francisco.
Last Updated on March 16, 2026 by Nick Ross
