What we’re about
Welcome to the Real-Time Data Lakes and AI meetup—a vibrant community at the frontier of real-time analytics, cloud-scale data lakes, and AI innovation. We're here to unpack the challenges of bridging fast-moving databases with scalable storage formats, from integrating Apache Iceberg to orchestrating lakehouse architectures with S3, Parquet, and ArrowStream. Expect deep dives into architecture, open table formats, and strategies for compute-storage decoupling. Whether you're debugging at midnight or dreaming of open data standards, this is your spot to learn, share, and shape the future.
Upcoming events (1)
See all- Open Source Analytics FestivalIBM Silicon Valley Laboratory, San Jose, CA
IMPORTANT: Please register here: https://lu.ma/wnsi90or . We look forward to seeing you there!
OSA Con is proud to announce an evening devoted to four of the most popular analytic open source technologies on the planet: ClickHouse®, Presto, Gluten, and StarRocks. We'll have presentations from community experts followed by a panel featuring audience questions to the presenters. There will be refreshments, drinks, and lots of time for networking with other database developers. Join us!
* OSA Con = Open Source Analytic Conference. The CfP for our main Conference in November is open now!## Presenters:
- Robert Hodges (Altinity; speaking on ClickHouse)
- Ron Kapoor (StarRocks / CelerData)
- Aditi Pandit (Presto / IBM)
- Binwei Yang (Apache Gluten / IBM)
- Panel Discussion: Ali LeClerc (IBM) will be host and interview the presenters.
## Description of the talks:
Building Cheap, Fast, Scalable Analytics with Open Source ClickHouse® by Robert Hodges, CEO @ Altinity
- ClickHouse is the go-to database for processing event streams and delivering user-facing SaaS analytics. This quick overview shows the major features that make ClickHouse so popular for real-time analytics and provides a jumping off point to build your own apps. We'll provide explicit guidance on when to reach for ClickHouse and how to get started. We will also demo new work from Altinity that adds separable compute and storage using shared Apache Iceberg tables. Users can take advantage of cheap object storage to process massive datasets without breaking the bank.
Real-Time Customer-Facing Analytics: From Pain to Production by Ron Kapoor, Developer Advocate @ CelerData
- Real-time customer-facing analytics drives growth and engagement—but only if it's fast, fresh, and reliable. In reality, many teams still struggle with:
- Queries that take seconds or even minutes when users expect instant results
- Latency spikes during peak traffic that break SLAs
- Expensive, fragile pre-computation pipelines
- Data freshness gaps that confuse users and undermine trust
- This talk explores the architectural patterns and open-source technologies that leading teams use to meet the demands of customer-facing workloads: sub-second latency, high concurrency, and real-time updates, without breaking the bank. We’ll share lessons from companies like Pinterest and Demandbase on how they tackled these challenges and what worked and what didn’t. Finally, we’ll look ahead at how open table formats and emerging AI agents are shaping the future of customer-facing analytics and how to build for today’s real-time needs while being ready for what’s coming next.
Pushing the Limits of Query Speed: Presto C++ in Action
Aditi Pandit, Software Engineer @ IBM- Presto C++ (aka Prestissimo) is a high-performance rewrite of the Presto engine in C++, designed to power interactive analytics at massive scale. In this session, Aditi Pandit will walk through recent performance optimizations in Presto C++, including memory and compute efficiency. Backed by benchmarks and real-world results, this talk will show how Presto is closing the gap between open-source flexibility and high performance.
Accelerating Spark Workloads with Apache Gluten and Velox
Binwei Yang, Software Engineer @ IBM- Apache Gluten unlocks native query acceleration for Spark by replacing the JVM-based execution engine with a C++ backend powered by Velox. In this session, Binwei Yang will share how Gluten is delivering up to 6x faster performance on key workloads along with some benchmarks. Get a look at recent optimizations in vectorization, I/O, and operator execution and why native engines are reshaping the future of Spark performance.
