Keebo

Keebo, founded in 2019, operates a data learning platform designed to address the escalating costs and performance challenges of cloud data warehouses for enterprises. The company was co-founded by CEO Barzan Mozafari and Yongjoo Park. Mozafari's extensive background as an Associate Professor at the University of Michigan and his 15 years of award-winning research at the intersection of machine learning and database systems at institutions like MIT and UCLA form the bedrock of Keebo's technology. This academic journey provided him with firsthand insight into the manual burdens overwhelming data teams, inspiring the creation of an automated solution.

The company's core business revolves around automating the optimization of cloud data warehouses, particularly for platforms like Snowflake and Databricks. Keebo targets data engineering and analytics teams within enterprises, aiming to free them from the time-consuming and manual tasks of managing data infrastructure. Its business model is value-aligned, often charging clients a percentage of the savings they achieve by using the platform, though a flat-fee option is also available. This approach allows clients like Allbirds, TUI, and Barstool Sports to reduce cloud warehousing costs by an average of 30-60%.

Keebo's platform utilizes patented, AI-driven algorithms and reinforcement learning to continuously monitor and optimize warehouse operations without human intervention. A key feature, Warehouse Optimization, dynamically adjusts resources by analyzing telemetry metadata—not sensitive customer data—to predict usage patterns and right-size warehouse configurations in real-time. This ensures that resources are scaled up during demand spikes and down during lulls to prevent overspending. The platform also includes "Performance Guardrails" that allow users to set specific parameters for query latency and queue times, ensuring that cost-saving measures do not compromise performance SLAs. Furthermore, its Intelligent Query Routing feature directs queries to the most cost-efficient warehouse, further reducing compute costs. The entire system is designed for quick implementation, integrating into a client's existing stack without requiring data migration or code changes.

Keywords: data learning, cloud cost optimization, warehouse optimization, Snowflake optimization, Databricks, query acceleration, data engineering automation, reinforcement learning, data analytics, FinOps, Barzan Mozafari, cloud infrastructure, AI-powered optimization, data warehouse automation, performance guardrails, query routing, metadata analysis, cloud spend management, automated resource management, data platform ROI, enterprise data systems