Next-Gen Data LakeHouse: The Iceberg Advantage
Twingo's Next-Gen Data LakeHouse: The Iceberg Advantage event brings together leading experts to explore all about Apache Iceberg, from metadata analysis with agentic AI, to performance optimization at petabyte scale, to enterprise ready extensions like LinkedIn’s OpenHouse, offering a focused look at the strategies and engineering practices driving modern data platforms.
AGENDA:
09:00 - 09:50 - Breakfast & Networking
09:50 - 10:00 - Opening Remarks
🔹Golan Nahum, CEO, Twingo
10:00 - 10:30 - Rebuilding Data for AI
This session explores Apache Iceberg’s metadata landscape using agentic AI alongside Spark and Trino, demonstrating how different specification versions, including V3, influence performance and analytical capabilities. Attendees will see how AI can generate meaningful insights from Iceberg metadata and gain a practical view of what succeeds and what fails when handling large scale and standard ingestion workloads. The session provides a grounded perspective on best practices, pitfalls, and lessons learned from real implementations.
🔹Aviv Erez, Data & AI Evangelist
10:30 - 11:00 - Optimizing Iceberg at Petabyte Scale: How Compaction Strategies Transform Data Lake Performance
This session highlights the architectural importance of compaction in Apache Iceberg and how it preserves high query performance as data volumes reach petabyte scale. It explains how binpack, sort, and z-order strategies reorganize data files to reduce scan overhead and improve filter efficiency, turning fragmented tables into optimized structures that accelerate analytics. Through production examples and benchmark results, the session demonstrates how effective file organization influences compute costs and query latency, and provides practical guidance on when to apply each compaction method, how to tune it as datasets grow, and which metrics best reflect ongoing performance gains.
🔹 Ran Reichman, Co-Founder & CEO, Flarion
11:00 - 11:10 - Coffee Break
11:10 - 11:40 - Let’s Open the House for Iceberg
Apache Iceberg is transforming how modern data lakes are built, offering powerful features like time travel and ACID transactions at petabyte scale. While it solves many challenges in data storage and querying, adopting Iceberg at scale in production still leaves some open questions. This session explores where Iceberg excels, where its limitations emerge, and how LinkedIn’s OpenHouse, an open source framework built on top of Iceberg, helps fill those gaps by adding structure around dataset ownership, versioning, governance, and self-serve operations. Attendees will gain a clear view of how Iceberg functions within the modern data stack and how OpenHouse can extend its capabilities to support a more enterprise-ready and scalable lakehouse.
🔹Asaf Halili, Staff Software Engineer, LinkedIn
11:40 - 12:10 - Optimizing Analytics on Apache Iceberg: Advanced Layout Tuning and SingleStore Serving Patterns
This session explores advanced physical design techniques such as partitioning, clustering, file sizing, and ordering strategies that align Iceberg tables with real query patterns and strict SLAs. It also examines the performance limits of native Iceberg combined with common query engines in scenarios like high concurrency, low-latency BI, operational analytics, and product-facing APIs. Finally, the session introduces how SingleStore can be layered as a serving tier on top of Iceberg, covering ingestion models, indexing methodologies, and reference architectures that unlock speed and efficiency for production workloads.
🔹Ilya Gulman, CTO, Twingo
12:10 - 12:30 - Experts Panel: Aviv Erez, Asaf Halili, Ran Reichman, Ilya Gulman.
12:30 - 13:00 - Lunch