Sep 22, 2026 · San Francisco, CA

Rows & Columns Summit

A practitioner-first conference on the architecture question that won't go away: OLTP and OLAP, together or apart? Real systems, real tradeoffs, real numbers — from engineers who've lived them.

Rows and Columns Summit logo
Date September 22
Venue Contemporary Jewish Museum
Format Single-track · 10am–7pm
Register for Event See the program
400+
Curated practitioners
1
Track · no breakouts
1 day
Full day + happy hour
$0
Admission · free to attend
The thesis

The Great OLTP–OLAP
Convergence

The convergence of OLTP and OLAP is a generational architectural shift, not a niche trend. As organizations push toward real-time analytics, AI, and operational intelligence, transactional and analytical workloads increasingly need to work together. But there is no single path forward. Hybrid architectures, open table formats, CDC, purpose-built systems, and new database paradigms each offer different approaches, strengths, and tradeoffs. This event brings together data practitioners, architects, vendors, and researchers to discuss the evolving world of transactional and analytical systems. Join us to debate architectures, share real-world experiences, and explore the ideas shaping the future of data systems in the era of AI.

September 22, 2026

The Program

Single-track from open to close.

9:00am
Doors Open
Registration & Coffee
Check in, grab a coffee, and settle in before the day kicks off.
10:00am
Keynote
Opening Keynote — Andy Pavlo
Setting the stage for the day.
10:45am
Panel
The Convergence Debate
Vendor leaders from across the ecosystem in conversation.
12:00pm
Lunch
Lunch on the Outdoor Terrace
A wonderful hour to network and get to know fellow attendees, plus good food.
1:30pm
Talk
Hannes Mühleisen
Details to be announced.
2:00pm
Case Study
ShipScience
ShipScience audits parcel invoices and recovers carrier claims for e-commerce and Fortune 500 shippers — 200M shipments a year and over 1B tracking events. Their production MySQL database is the system of record, and it's messy by necessity: shipment data scraped from carrier portals by browser agents, tracking feeds over API, webhook, and SFTP, emailed invoices, OMS/ERP syncs from Shopify and Manhattan, and everything humans create in the UI — all while staying fast for the app. When they started building agents and ML models on that data, they weren't willing to let those workloads compete with user traffic, so they moved to streaming row-level CDC out of production with Artie, landing in MotherDuck within minutes across ~30 tables and ~3B row updates a month. This session covers why streaming CDC beat scheduled extracts (with implementation lessons), how Claude plus the MotherDuck MCP over application logs became an unplanned win that changed how their engineers debug, and ActiveAudit in production — what agentic auditing catches that rules-based auditing missed, and what it recovered for customers.
2:30pm
Case Study
Rooftstock
Roofstock needed to serve historical, data-heavy reports inside their application without turning every user request into an expensive analytical query — or maintaining duplicate reporting logic across product and warehouse. This session tells the engineering story of moving that workload to a continuously updated analytical serving layer: the architecture before and after, and the production decisions that made it work — deciding which data needed to be fresh, designing transformations that serve both application and analytical use cases, handling schema changes and backfills, validating accuracy during cutover, and managing the cost of an always-on workload. The result: report rendering dropped from minutes to seconds. Attendees will leave with a decision framework for recognizing when operational reporting has outgrown query tuning, a reference architecture for decoupling reports from operational data stores, and practical guidance for evaluating whether this design fits their own systems.
3:00pm
Lightning Talks
Speaker TBA
Details to be announced.
3:30pm
Break
Afternoon Break
Stretch your legs, refuel, and visit the sponsor tables.
4:00pm
Talk
Russell Spritzer
We’re spending the day here talking about OLAP and OLTP data systems. Hybrid data systems, operational analytics databases, Postgres-based lakehouses, and streaming platforms. One thing that is clear, the industry is converging on Apache Iceberg as their shared data layer. This talk traces my storied past with OLAP/OLTP connectors, how Apache Iceberg has finally provided the ACID compatible table format we need for interoperability, and how OLTP use cases are driving the next generation of updates to Apache Iceberg V4. The open source community is growing, join us and learn how you can be a part of the next generation of hybrid systems.
4:30pm
Talk
Bonnie Xu
OpenAI’s data platform spans 70,000 datasets, making it difficult to find the right source and understand what its numbers mean. This talk explores how an AI data analyst navigates that complexity: searching metadata and lineage, drawing on institutional knowledge, querying relevant data, and showing the evidence behind its answers. We’ll share lessons in making data exploration at this scale both faster and verifiable.
5:00pm
Closing Keynote
Fatma Ozcan
Wrapping up the day.
5:30pm
Happy Hour
Evening Reception
Outdoor terrace and indoor sponsor tables. Great way to wrap up the day.
DET Community Booth
All day at Rows & Columns Open to All

Throughout the day, you'll be able to stop by the booth to:

Networking Sessions
Meet fellow data engineers during the event, start meaningful conversations, and make the most of your Rows & Columns experience.
1:1 Mentorship Sessions
Interested in 1:1 mentorship? Stop by the DET booth to check mentor availability and register on the spot.
Organising committee

The Board

Rows & Columns is organized by a small independent board of three including practitioners and community leaders with deep roots in data infrastructure. Backed by ClickHouse, with an independent program committee and blind CFP review.

Elizabeth image
Elizabeth Christensen
Snowflake
Open source developer advocate at Snowflake working on the Postgres and open lakehouse projects. She is PostgreSQL community contributor, board member a the US PostgreSQL Association, host of Postgres Meetup for All, and PostGIS Day.
shanes image
Shane Glass
Google
Shane is a senior developer advocate and open source strategist for Google, working across numerous open-source analytics communities and serving the Jupyter Foundation Governing Board. A lifelong learner with a Master's in Data Analytics. He loves connecting people, Pittsburgh sports, and finding the perfect donut.
Zoes image
Zoe Steinkamp
ClickHouse
Senior Developer Advocate at ClickHouse focused on amazing open source databases and in person community events. Active agent of chaos across the data community with a background in problem solving and engineering.
The space

The Venue

We're hosting Rows & Columns Summit at the Contemporary Jewish Museum in the heart of downtown San Francisco. Designed by Daniel Libeskind, the building is a striking blend of a restored 1907 power substation and a bold modern steel cube — a fitting backdrop for a conference about bridging old and new architectures. The venue features a 400-seat auditorium, a beautiful outdoor terrace for lunch and happy hour, and indoor gallery space for sponsor tables and networking.

736 Mission Street, San Francisco, CA 94105 · thecjm.org