🚀 DataHub 2.3 is live, with The Context module in public beta! This release closes the loop on agent accuracy with deeper evals, a feedback loop capturing what agents get wrong in production, and metrics fully wired into lineage. Get the full breakdown: ordnl.link/512mZsU

The knowledge that runs a business lives scattered across glossaries, metrics, lineage, and a few people's heads. At CONTEXT 2026, ReKnew CTO Les D'Souza makes the case for treating it as strategic infrastructure. Free, virtual, Nov 4: datahub.com/context/?utm_cam…

"Active customer" has three definitions across your company, and your agent picks whichever the schema points to. A semantic layer settles it. Take a look at where it fits in context management. ▶️

We're building the open-source standard for context management, the layer that decides whether an AI agent can be trusted in production. We're hiring across engineering, product, and go-to-market. See the open roles: datahub.com/datahub-careers/…

Context gets more hype and less definition than almost anything in AI. At CONTEXT 2026, BARC VP of Research Kevin Petrie lays out the pillars of context engineering, then joins DataHub co-founder and CEO Swaroop Jagadish on building AI trust. Free, virtual, Nov 4: ordnl.link/RKYXsJO

When the main consumer of your data is an AI agent, governance can't be an afterthought. Jeff Miller of @BlackRock joins DataHub's @Shirshanka at CONTEXT 2026 to break down how BlackRock makes data agent-ready. See you there: ordnl.link/DCBYDRj

Trustworthy data is what AI in production runs on. Super turned a wall of quality checks into exactly that on Snowflake, Airflow, and DataHub. Context management in practice. Catch Nikola Kljajo at CONTEXT 2026 for the how.

A typical data quality tool marks everything critical, so teams stop reading the alerts. 😬 Our Open & Wildcard hackathon winner, Paracelsus by Lutfiya Miller & Chris Müller, re-ranks @Snowflake and Postgres findings by real risk using a model from toxicology. ▶️ piped.video/wt2gXQ1ibG8

We finally got the whole DataHub team together in one timezone for two days in Chicago! 🎉 H1 planning, a marketing hackathon on our dream Claude workflows, dinner on the river, and a very competitive dart tournament. Grateful for this small but mighty crew. 🚀

Your newest hires this year may be agents, not people, and like any new hire, they're lost without context. At CONTEXT 2026, Prajakta Damle, Senior Director of Product Management at @GoogleCloud, sits down with DataHub co-founder and CTO @Shirshanka for a fireside chat on onboarding agents into the enterprise. They get into how to bring context into one place and keep it current as agents take on more work, and why open-source community investment is where the future of context gets built. 📍CONTEXT 2026 - November 4: ordnl.link/nL9wb6T

Your agents guess in production and never say they're doing so. DataHub Cloud grounds every answer in validated context, and Context Feedback catches the misses and routes them to the right person before they repeat.

Clean data and a strong model still leave agents confidently wrong. The missing piece is reliable context at query time, and here are the 10 principles behind a platform that delivers it. ordnl.link/Cs0F9fC

How do you keep data trustworthy when agents run your data at machine speed? Track 2 at CONTEXT 2026 digs into exactly that. Catch experts on domain knowledge as infrastructure & when shipped and maintained. Plus scaling AI readiness in DataHub. 📍Free, virtual, November 4: ordnl.link/PHe9PyA

A lot of AI projects never reach production, and context is usually why. Our new guide lays out the 10 principles of an enterprise context platform: ordnl.link/5aluzmR

"Our engineers would've rather taken on-call than do release merges." So DataHub built an agent for exactly that, human sign-off included, moving releases from ~6 weeks to daily. @Shirshanka on @Snowflake's Inside the CTO Circle w/ @ShrutiBhat & @jtroyer: piped.video/watch?v=4h5S3i1S…

Our community recently crossed 16,000+ members across 3,000+ organizations. A testament to the open source community—practitioners comparing notes and sharing what they've learned about running context management in production. If you're building toward AI, we'd love to have you: ordnl.link/vq7DlfC

Your analytics agent is in production and some of its answers are quietly wrong. Trace them back and you usually land on a context problem. CONTEXT 2026 is where teams share how they fix it. Free, virtual, November 4: hubs.la/Q04z2nWl0

DataHub's first German-language user meetup is happening, and it's open to the whole DACH community. ⚒️ It's happening online, so practitioners can join from wherever they are. It's a chance for German-speaking DataHub users to compare notes on running DataHub in production and where context management fits in their stack, in their own language. Register 👉️ ordnl.link/FlyEInn

The show goes on! DataHub has been on the @Snowflake World Tour this fall and we've got two more stops to go! It's been a blast in Stockholm and London and we can't wait to see you at the next stops. 🏙️ Chicago — Today! October 5 🗽 New York City — October 15 We look forward to connecting with the Snowflake community in these iconic cities. Get ready to talk all things data and context!

A vendor changes one column, its max shifts from 6 to 7, and the quality checks stay green. 😬 Meanwhile a production model racks up $90,322 of errors in a month. Our hackathon's production ML winner, Culprit by Jonathan Andrei, catches exactly this. ▶️ ordnl.link/VYSTFRr