MotherDuck

Tomorrow at 9am Pacific, MotherDuck co-founder Ryan Boyd 🐤 will join Ryan Dolley in a live session of Super Data Brothers to talk about all things MotherDuck, DuckDB, DuckLake, agentic analytics and more. Join us: https://lnkd.in/gJ_szqAW

At MotherDuck, we want to make it easy for our customers to build Agentic Applications on top of our Database. With that in mind, we have written a guide on proven patterns for building these, as well as built functions with the AI engineer in mind. One of those functions is Use __MD_FILTER_TABLES(), which allows fuzzy search across tables in the current database. These types of functions directly in the database make it faster to give the context needed to your agent, and decrease the wait time for the end user when seeking answers. Take a look at our guide to get started with building your own agents! https://lnkd.in/gNzX3Swm

2 weeks!! Small Data SF is in 2 weeks!! Some of our hands-on workshops only have a couple seats left in them, so hurry up and register to get your choice. Here are some great workshops with availability: * bem's co-founder and CTO will teach how to take unstructured data (like PNGs, PDFs, etc) and query it in your data warehouse. * Hex is running a data app hackathon with real prizes. Who doesn't want to build a data app? * Estuary will teach you how to get real-time streaming data into your data warehouse with ease. Book your seat while they last! https://lnkd.in/g_abRnUp

Thanks to our friends at bem for sharing! Excited to have this workshop on unstructured data (think PNGs, PDFs, etc) into MotherDuck, led by Upal Saha, Co-founder and CTO at bem.

Gold layer analytics stuck in the stone age? Join MotherDuck and Artefact for real-world benchmarks showing how DuckDB transforms performance in the gold layer—from slow dashboards to sub-second queries. We'll cover: 🦆 Benchmarking results for gold layer workloads with DuckDB 🦆 Why OLTP databases bottleneck your analytics 🦆 How DuckDB's vectorized execution changes the game 🦆 Practical patterns for performant architecture Register now:

Jordan Tigani kicked off the small data movement with his "Big Data is Dead" blog post. He shared why he's excited about hosting and speaking at the "Small Data SF" conference coming up in 2 weeks.

Increasingly we're seeing AI agents querying MotherDuck data warehouses. This isn't just the typical one-off 'text to SQL' use cases (which we support, but are still hard for humans to build confidence in), but rather AI agents that seek to have a really strong understanding of the underlying data schema and build their own semantic layers through iterative SQL queries. My co-founder Jordan Tigani recently gave a talk at AI Native on why AI needs the help of analytics databases for their work -- and I'll be talking at the upcoming AI By the Bay Conference on a similar topic. AI is magical, but tends to be pretty bad at facts and especially bad at large scale aggregation of facts. While they can benefit from full semantic layers like Cube or those built into BI tools like Omni, you can also provide 'poor mans' semantic layers using simple COMMENT ON <TABLE|COLUMN> in DuckDB as well as creating constrained views that helps make AI agents better informed. Of course, you also have a lot to worry about access control and potential runaway costs. MotherDuck's hypertenancy model, where each user gets their own DuckDB-based duckling and constrained access, makes it really easy and efficient to build sandboxes for AI to play in. More info in our new guide on Building Analytics Agents with MotherDuck, written by Till Döhmen, our AI/ML lead. It even includes an example agent with the OpenAI Agents SDK and how to use Claude Desktop with the MotherDuck MCP connector: https://lnkd.in/dDvV9j8y

With the AI hype, many companies forget there’s a gap between what they want and where they actually are in data maturity. You need solid data foundations before stacking AI on top. ft. Salma Bakouk from Sifflet

That's a wrap on Coalesce 2025!! Thanks to everyone who stopped by our booth! Here are some of our highlights: 🎁 Sharing our amazing swag, thoughtfully curated by Carly Spoljaric 🦆 Alex Monahan and Jacob Matson's DuckLake talk and demo 🛟 Donating to International Bird Rescue on behalf of our scratch and win winners to rescue ducks If you missed us, we'll be at three more conferences in November: - Small Data SF, Nov 4-5 (https://lnkd.in/eGb2_gbH) - Data in the D Conference, Nov 8 (https://lnkd.in/ga6C7EXz) - AI By the Bay, Nov 17-19 (https://ai.bythebay.io/)

8 Hands-on Workshops hosted by experts on Nov 4th as part of our SF conference focused on fast, easy-to-use, cost-efficient data tooling. 5 for data engineers, and 3 for AI, BI, Viz engineers/analytics. We have you covered! 1. Learn how to build a serverless lakehouse with the MotherDuck team. 2. Craft composable data pipelines with Dagster Labs 3. Pythonic data ingestion and transformation with dltHub 4. Build a real-time data stream into MotherDuck with Estuary 5. Transform unstructured data (eg PNGs, PDFs) for SQL querying with bem 6. Build AI app using semantic models with Omni 7. Hack for prizes at the Hex data app hackathon using quirky data 8. Build an autonomous data science agent with Together AI Learn more and register for Small Data SF soon while there are still spots available for these workshops!