Numbers Station AI

Business intelligence tools are limited. Analytics agents are the future. And it's just the beginning. 🚀 Check out our demo below! Full release on our blog: https://lnkd.in/g3caKRkM

Speak in your business's language and extract knowledge from your existing data catalog through a direct integration with your warehouse. Dive into engineer Peng Xu's blog on how we easily unify Tableau workbooks (and many other warehouses) to power Numbers Station's semantic catalog. 😌 Effortless extraction through seamless warehouse integrations 👩🏻💻 Parser agents that intuitively processes and sifts through catalog information 🧠 LLM robustness with first-hand understanding of your proprietary business speak https://lnkd.in/gwG9596d

Recruiting is THE most leveraged thing founders do. If you recruit well, you can do many other things poorly and still thrive. The good news is that who and how you hire are the most predictive and controllable inputs to your company's success. The bad news is that many founders approach recruiting as a series of transactions rather than as a strategic pillar of company building. As the company grows, founders might be tempted to hire a recruiter or outsource parts of the recruiting process to team members or agencies. Don’t succumb to this notion. Now is the time to triple down on founder-led recruiting, not hand it off or staff it out!! Read more on founder-led recruiting & subscribe to Hooked on Substack: https://lnkd.in/gsTBT7hH Thanks, Christopher Aberger (Numbers Station AI) for your insights! Shannon Anderson

“Don’t make me think!” Engineer Tywen Kelly introduces 'Optioneering'—the basic principle for how to wrangle LLMs into an ergonomic interface. Users should not be expected to be engineers, nor should they be pandered to. Optioneering strikes the right balance of control & guidance. ↓ ✍🏼 Great UX reduces a user's cognitive friction of getting started 🧠 Bring humans into the loop by presenting the LLM's step-by-step reasoning process 📝 Generate templates users can build upon to generate effective outputs https://lnkd.in/gX5Ux8dA

Early results show that Meadow is within 3.6 points of state-of-the-art on the Spider text-to-SQL benchmark and realizes lifts of up to 8.8 points over simple, zero-shot text-to-SQL prompts. 🚀 🚀 🚀

We wish you were there... ☘️ The Numbers Station team put remote working on pause for our bi-annual offsite—this time in Seattle, Washington! A wonderfully diverse group for a diverse set of activities—from team workshops, breakout sessions, and guest speakers to hiking, sailing, and a (surprisingly competitive) night of pool 🎱

The Databricks Data+AI Summit festivities continue! 💪🏼 Find Chief Scientist Ines Chami & ML Architect Laurel Orr at the Compound AI Systems Workshop tomorrow (11:30am - 6pm PDT) for a lightning talk and poster on our newest OSS, Meadow! 🪴 https://lnkd.in/gpRPuZ8P

We're kicking off our first talk on Meadow at the MosaicX Meetup at Databricks' Data+AI Summit tonight. 🎉 Hear ML Architect Laurel Orr deep dive into how we've built the simplest LLM framework for data tasks, along with CEO Christopher Aberger in attendance. We'd love for you to join us!

Numbers Station Labs is extremely excited to release Meadow, the first and simplest agentic framework built for complete data workflows. Meadow is fully open source! 🪴 Early results show that Meadow is within 3.6 points of state-of-the-art on the Spider text-to-SQL benchmark and realizes lifts of up to 8.8 points over simple, zero-shot text-to-SQL prompts. 💪🏼↓ Github: https://lnkd.in/grdRZG-P https://lnkd.in/gQ9fxfrg

We're so excited to share an early preview of Meadow—the first multi-agent framework for Text2SQL—to the Databricks #MosaicX Meetup next week! 💡 Discover the reasons why existing LLMs fail at translating business questions into SQL, and our solution to truly deploying data querying beyond technical teams. 📆 June 10, 2024; 8:35-8:50am 📍 Moscone Center South, 2nd Fl, MPT Stage