As Apache Doris 2025 MVP, core contributor to Doris-MCP, and Co-Founder of Yiwen Data Agent, I will leverage my experience supporting over 500 leading enterprises in their data warehouse transformation within the Apache Doris community, combined with hands-on implementation expertise in the AI4Data field, to share an end-to-end solution for building a zero-illusion AI-assisted decision-making platform on Apache Doris.

This session will dispel industry misconceptions about AI4Data, focus on the real-world decision-making needs of enterprises, and deliver a replicable, validated practice framework for AI + data engine integration to the Apache Doris community, covering the full lifecycle from technical architecture design to production deployment:

1. Zero-Illusion Technical Foundation: From “SQL Guessing” to “Deterministic Generation” Our innovative NL2LF2SQL dual-engine architecture enables step-by-step conversion from natural language to logical form and finally to SQL, completely eliminating the hallucination problem in large model data retrieval and achieving 100% data accuracy aligned with financial audit standards. Powered by Apache Doris’ real-time unified architecture, AI-driven decision-making is backed by millisecond-response computing capabilities.
2. Production-Grade Implementation Practice: From Prototype to Enterprise-Scale Deployment We will share best practices for building an AI4Data platform on Apache Doris, including how to directly reuse DWD layer detailed data and eliminate the need for complex middle-layer construction, enabling enterprises to build intelligent decision-making platforms at a low cost with reduced implementation complexity.
3. Business Value Closed-Loop: From “Data Retrieval Tool” to “Strategic Decision Advisor” We will demonstrate how to extend AI capabilities beyond basic data retrieval to cover the full value chain of attribution analysis, business insight generation, and actionable decision support. Through our “dual-brain collaboration” digital employee model, AI not only retrieves data accurately but also provides actionable decision recommendations for businesses, truly unlocking the business value of AI4Data.

This session will equip Apache Doris community users with an AI4Data implementation solution that balances technical depth and practical value, helping enterprises build a trustworthy, efficient, and value-driven intelligent decision-making system in the era of large models.

Speakers:

奕嘉 苏: Yiwen Data Co-Founder、Apache Doris Committer、PowerData Sponsor

Yiwen Data Co-Founder, Apache Doris Committer, Apache Doris Community Evangelist, Doris-MCP Contributor, PowerData Community Initiator, has helped dozens of enterprises complete the production and implementation of Data Agent and ChatBI projects, and is committed to building a zero-illusion auxiliary decision-making platform, and has also assisted hundreds of companies in the Apache Doris community to complete real-time data warehouse construction and optimization evolution.