Small Data SF
Two Full Days, One All-Access Pass: $379
Nov. 04-05 San FranciscoNov. 04-05 2025 San Francisco
Make Big Data
And AI
Feel Small
Local-First. Future-Ready.Local-First. Cloud-Free. Future-Ready.
Something small is happening
Small Data SF returns to San Francisco on November 4-5, 2025, bringing together developers and data practitioners for two days of workshops and talks that challenge the "bigger is always better" mentality in data and AI. Join us to explore the philosophy that's reshaping how we think about data processing! Day 1 (November 4): Hands-on workshops where you'll learn practical techniques for efficient data processing, local-first development, and building AI applications that don't require a cluster to run. Day 2 (November 5): A full day of talks from industry leaders who are redefining what's possible when you optimize for simplicity, speed, and developer experience rather than theoretical scale.
Small Data isn't just about datasets that fit on a single machine—it's about efficiency in making big data feel small, processing data in smaller pieces before it gets too big, embracing local-first development, and choosing simplicity over unnecessary scale. With computers now one hundred times more powerful than during the early Big Data hype, it's time to focus on the underrated joys and untapped potential of laptops, cloud machines, and edge computing. This same philosophy extends to Small AI, where smaller, optimized models offer speed, flexibility, and practical deployment options without the overhead of massive infrastructure. Because sometimes the question isn't "How big is your data?" but "How efficiently can you process it?"
The Small Data (& AI) Movement is already underway. Are you in?
Highlights of 2024
Small Data. Huge Impact.
Manifesto
It’s time to think small.
We believe in the Simple Joys of Small Data.
It’s finally possible to hold Big Data in the palm of your hand.
Is that Big Data or Small Data? That’s for you to decide.
Local-first development is more powerful than you think.
Modern hardware is a beast.
Small models drive huge impact.
Small data, smart AI.
Think small, develop locally, ship joyfully.
More Data ≠ Better Results.
Our data is subject to the law of diminishing returns.
Most data we use is recent data.
Recent data is the most valuable slice of our data.
Bigger data has an opportunity cost: Time.
Machine Learning on large data is expensive, and small models deliver huge impact.
Less is more.
Single Machines are Efficient and Powerful.
Simplicity is scalable: chaos and complexity are not.
Scale-up workflows are highly efficient, with 400X more memory than 10 years ago.
Let’s focus on delightful user experiences instead of fumbling with complex infrastructure.
We can save distributed compute for when we really need it most.
Single machines rule.
Developing Locally Just Works.
Local dev delivers instant results, while local + edge computing unlock powerful new experiences.
Ditch the distributed complexity when you don’t need it.
Not everything we build should require inefficient containers or a round-trip to the cloud.
Hardware and laptops are 100X more powerful, yet we leave this compute dormant and unused.
We believe in developing locally and shipping to prod with the same software.
Together, we can end the cloud hangover and support local analytics.
Small data and AI is more valuable than you think.
2025 Speakers
Agenda
Registration
The Great Data Engineering Reset: From Pipelines to Agents and Beyond
Joe ReisDeepLearning.AI
Projection Pushdown vs Predicate Pushdown: Rethinking Query Efficiency
Adi PolakConfluent
Panel: Don't duck up the Numbers: Where AI Hype Meets BI Reality
The end of small data
Benn Stancil
Lunch
From 1 parameter model to 100 billion parameter models. How progress on one benefits all
Ravin KumarGoogle DeepMind
Building Distributed DuckDB Processing for Lakes
George FraserFivetran
Small Data vs Big Hits - Using Local Data Tools to Map the Songwriting Industry
Sam AlexanderExtended Play
Small data, big features: we are rewriting SQLite!
Glauber CostaTurso
Doing More with our Data at DoSomething.Org: a Nonprofit’s Small Data Story
Sahil GuptaDoSomething
Explore23: Web application for exploration of a large genomic research cohort
Teague Sterling23andMe
Better Data, Smaller Models, Bigger Impact
Shelby HeineckeSalesforce
4 Lessons Learned the Hard Way and How Thinking Small Saved the Day
Scott HainesBuf
Closing Panel
Networking Reception
[Check out the best moments]
Small Data is Bigger than Ever
BUILD BIGGER
WITH SMALL
DATA AND AI
NOVEMBER
04-05
2025
Local-First. Cloud-Free. Future-Ready.
From the community
2025 edition Presented By
2024 edition was Presented By
Small Data. Huge Impact.
Are you ready for this year?