Starburst logo

Agenda

Trino Day – October 22, 2025

10:05am - 10:35am ET

Beyond the Queries: Insights from Trino Experts

Join us for Beyond the Queries: Insights from Trino Experts, a panel discussion with some of the leading contributors to Trino. In this session, the team will share their perspectives on some of Trino’s latest innovations throughout the last year, as well as recommendations from the experts.  Whether you’re an engineer, data practitioner, or simply curious about the inner workings of Trino, this session offers a unique opportunity to go beyond queries and understand the ideas, decisions, and expertise that drive one of the most powerful data query engines in use today.

10:45am ET

Migrating a decade of Redshift usages to Trino at Quora

Quora migrated its decade-long Redshift analytics workloads to Trino on S3 to cut costs and simplify operations. The move reduced system complexity, improved scalability, and delivered ~50% cost savings while maintaining reliability. Though some query optimizations were needed, the transition resulted in a more flexible and manageable data platform. Hear firsthand from Gabriel Fernandes de Oliveira and Eliza Pan as the duo go over the migration process, challenges and learnings.

Speaker(s):

11:15am ET

Banco Data Without Borders: Banco Inter’s Federated Chargeback Model

Banco Inter, Brazil’s first fully digital bank, struggled with costly, slow, and siloed data systems that limited scale and innovation. To unify access across diverse sources and reduce query costs, the team adopted Starburst Enterprise as a federated query engine. Join Vitor Mattioli and Pedro Almeida as they detail their unique chargeback strategy which splits costs across 25 different data domains in three different categories: storage, processing machines, and data consumption.

Speaker(s):

Author biopic for Vitor Mattioli (Banco Inter)
12:11pm ET

Inside Our Journey to a Unified, Scalable Query Layer with Trino

At PlaySimple Games, millions of players generate terabytes of data daily, demanding fast, reliable, and cost-efficient analytics. Initially, we relied on Redshift for ad-hoc queries and Spark for pipelines.

The Challenge: Redshift and Spark created duplication, rising costs, and scaling bottlenecks. We migrated to Trino to unify our query layer and handle both ad-hoc and scheduled analytics at scale. Along the way, we split workloads into optimised clusters, tuned performance. The results were transformative: query volumes grew 4x (from 4K to 15K per week), P95 latency dropped from 240s to under 90s, all while delivering 35–40% annual cost savings. Today, Trino powers 50K+ weekly queries backed by 10K+ Delta tables across teams, enabling faster experimentation and insights without compromising efficiency. This talk will share our migration journey, performance tuning lessons, and cost optimisation strategies—takeaways for any team scaling Trino in production.

Speaker(s):

12:33pm ET

Conquering Data Skew in Trino: Enterprise-Scale Query Optimization Strategies.

Data skew remains a critical performance bottleneck in distributed query engines, affecting 30-40% of large-scale analytics workloads and causing severe resource imbalances across Trino clusters. This presentation addresses the fundamental challenge of uneven data distribution that forces certain workers to process gigabytes while others remain idle, transforming queries that should complete in minutes into hour-long operations.
Drawing from production deployments at Roku and other data-driven enterprises, we explore three battle-tested optimization techniques specifically adapted for Trino environments. We demonstrate dynamic repartitioning strategies that leverage Trino’s distributed architecture to achieve 20-40% performance improvements for moderately skewed datasets. For extreme skew scenarios involving hot keys in join operations, we introduce salting techniques that artificially increase key cardinality, enabling Trino to distribute processing loads effectively across all available workers.
The session culminates with broadcast join optimizations, particularly powerful for star schema queries common in data lakehouse architectures. By utilizing Trino’s broadcast capabilities for dimension tables, we eliminate expensive shuffle operations and dramatically improve query performance. Throughout, we emphasize practical implementation considerations including optimal partition sizes, memory configuration for Trino workers, and monitoring approaches using Trino’s built-in statistics. Attendees will gain actionable insights for identifying skew patterns in their query plans and implementing solutions that reduce processing times while optimizing cluster resource utilization in their Trino deployments.

Speaker(s):

Virtual Day – October 23, 2025

10:15am ET

How to sell Trino to your boss

You already know that Trino is fast, powerful, and proven at scale. The real question is how to get your boss to sign off without getting stuck running clusters by hand. This session, featuring Starburst CEO Justin Borgman, whose engineering background drives a hands-on, product-first approach, lays out the blunt case for Trino. Learn how to sell Trino internally by framing the ROI your team can expect when you stop wasting cycles on ops. Trino delivers fast SQL queries at scale across any data source without data movement, and managed Trino gives you all of that power without the maintenance drag.

Speaker(s):

10:50am ET

SpotHero Spotlight

TBA

Speaker(s):

11:25am ET

Breaking Down Data Silos: On-Prem Lakehouses with Trino

Security Operation Centers (SoC) rely on a SIEM to ingest and alert on logs that indicate risks in their environment. SIEM vendors can charge massive amounts for what is essentially a database with canned queries! I recently got the chance to build a SoC from the ground up and I desperately wanted to avoid the pitfall of security vendor lock-in; so I decided to build my SoC around Starburst Galaxy and Apache Iceberg. After 10 months in I wanted to share my architecture and lessons learned for others looking reduce costs or even entirely escape the grasp of their SIEM vendor.

Speaker(s):

Save Your Seat!

Day 1: Trino Day — deep dives into real-world architectures, performance tuning, and integration best practices from engineers running Trino at scale. 

 

Day 2: AI+Datanova — see how Starburst is powering next-generation AI and analytics, and get an exclusive look at our upcoming product roadmap. Leave with proven strategies, fresh implementation ideas, and a clear path to getting more from your data stack.

 

AI & Datanova is your gateway to the future of data and AI innovation.
Whether you’re a data engineer, architect, analyst, or business leader, this is an event you won’t want to miss.

Secure your spot today — registration is free!

Register once to secure full access to both events!

Presented by