As data architectures evolve, maintaining separate silos for batch and streaming processing — the hallmark of Lambda architecture — has become increasingly costly and complex. How can we build a truly unified platform that delivers both real-time data freshness and high-performance analytics at scale?

In this talk, we go beyond theory and dive into the trenches of building a next-generation Unified Lakehouse centered around Apache Paimon. We will share battle-tested best practices and real-world implementation patterns, demonstrating how to architect a seamless data pipeline through deep ecosystem synergy:

Apache Flink + Paimon — Robust, low-latency real-time ingestion into a transactional data lake Apache Paimon — The core storage layer enabling ACID transactions, schema evolution, and unified batch-streaming reads StarRocks & Apache Spark + Paimon — Delivering exceptional interactive and batch query performance directly on the lake Apache Kyuubi — The unified, serverless SQL gateway tying it all together Apache Gravitino — Unified metadata management, enabling centralized metadata governance across engines and data sources Attendees will walk away with a proven methodology for building a unified Lakehouse around Paimon and other key components, along with practical guidance on component selection, integration, and production tuning across the ecosystem, as well as additional Data+AI application scenarios.

Speakers:

Zhoulong Liu: Senior Big Data Specialist, eclicktech

The head of the Big Data Department at eclicktech previously worked at Sohu Video and Tencent and is skilled in building big data platforms and related systems.