How SQLFlow Is Making AI as Simple as Writing a SQL Query

On May 6, Ant Group’s Vice‑CTO Hu Xi announced the open‑source machine‑learning tool SQLFlow, aiming to lower the AI adoption barrier so that engineers can invoke AI as easily as writing SQL.

SQLFlow combines the simplicity of SQL with the power of machine learning, allowing data engineers to build AI‑enabled applications without deep expertise in mathematics, statistics, or programming.

He Changhua, chief architect of computing and storage at Ant Group, leads the AI Infra team that developed SQLFlow. A Stanford Ph.D. with experience at Google and Airbnb, he also spearheads the next‑generation real‑time big‑data platform.

The platform seeks to fuse OLTP and OLAP in a fully real‑time system, enabling instantaneous transaction processing and analytics, which is crucial for scenarios such as financial risk control.

He envisions a “Big Data Base” that unifies data ingestion, real‑time model training, and instant decision‑making, effectively democratizing data and AI capabilities across industries.

“We want to make data and intelligence accessible so that anyone can develop applications without knowing the underlying details,” He says.

By integrating streaming, graph, and machine‑learning workloads, the new engine moves toward a truly unified, high‑efficiency data‑intelligence platform that can serve both Ant Group’s massive financial services and broader enterprise needs.

He believes that achieving this “full‑fusion” will position Ant Group at the forefront of global technology innovation.