CMO @CollateData, dad, husband, cyclist, and head banger.
Austin, Texas
Joined August 2008
- Tweets4.9K
- Following1K
- Followers977
- Likes3K
A highlight from my trip to London was meeting Angelita is person. She's built an amazing "AI virtuous cycle" where AI serves insights to people who validate and improve it.
More business users are getting trusted self-service answers than with any traditional BI tool!
@Scout24 runs 10+ real estate companies across Germany, Austria, and Spain on one AI context layer, built with Collate.
Live in production in under 10 weeks. One source of trusted context for people and agents.
🎥 Watch the full Scout24 story --> youtu.be/jQf8WvPHgSw?si=TNr5…
#AIforData #ContextLayer #CustomerStory #OpenMetadata
Steve Wooledge retweeted
🚨Happening tonight in San Francisco!
Open Lakehouse & AI Meetup, hosted with @Dremio and @Lensesio.
📆 Thursday, October 8 | 6–9 PM
📍 Hardwood Bar & Smokery, 680 8th St, San Francisco, CA 94103
🍕 Food & drinks provided
Hear from:
🎤 Sriharsha Chintalapani (CTO & Co-founder, Collate) - "Iceberg Made Tables Open. What Makes Context Open?"
🎤 Tun Shwe (Lenses) - "Understanding Agent Harnesses"
🎤 Alex Merced (Dremio) - "The Open Agentic Lakehouse: Building an Open Data and AI Ecosystem"
If you're in the city and want to talk open lakehouse, agentic AI, or just grab a bite with people building this stuff, come on by.
Spots are limited; save yours now 👉 buff.ly/JiT9qDh
#OpenLakehouse #DataEngineering #AI #Iceberg #DataGovernance #SanFrancisco #OpenMetadata #OpenLineage
Steve Wooledge retweeted
@JessicaTalisman has spent 25 years on the part of AI that nobody posts about.
Most AI deployments skip all of it, then wonder why the agent is confidently wrong.
Oct 14, 8am PT, with @d3fmacro.
buff.ly/ntz2Eqo
#Ontology #Semantics #OpenMetadata
ALT Agent-Ready Context Layer
Steve Wooledge retweeted
Incredible how her voice sounds much better as she gets older. Excellent artist.
Great perspective from @d3fmacro on the need for an agent harness on enterprise data and how it's built into @CollateData.
Claude Code works because of the harness around the model, not the model alone.
Enterprise data needs the same architecture.
buff.ly/VY3LiK9
#AgenticPlatform #AIAgents #OpenMetadata #Agentic #Harness #contextplatform
Steve Wooledge retweeted
Day 1 at @BigData_LDN: @suresh_m_s + @mikeferguson1 join @Scout24's Angelita Frozza Sanches on running AI agents in production on a trusted context platform. Add to your schedule here --> buff.ly/EEm1OEa
Day 2: @SanjMo, the analyst who wrote the book on it, signs copies at Stand K121, 10:30-11:30am.
#BigDataLDN #AIAgents #OpenMetadata #ContextPlatform
ALT Sanjeev Mohan book signing at the Collate stand, K121, Big Data LDN
My first time seeing them last night. Amazing show and great tribute to Frank Beard who was with us in spirit!
The Texas rock band turned its ACL Live performance into a celebration of its longtime drummer. statesman.com/entertainment/…
Steve Wooledge retweeted
"Sounds of Seattle" 1991, festival gratuito pagado por el ayuntamiento para publicitar el talento local. Pearl Jam cierran el concierto a tres días de publicar su primer álbum "Ten" y a un mes de que saliera el "Nevermind" de Nirvana que cambiaría la industria musical de un día para otro.
"Ten" no cuajó en un primer momento tras su lanzamiento y tuvieron que pasar ocho meses hasta que la fuerza de arrastre de "Smells Like Teen Spirit" —y el repentino interés hacia otras bandas grunges por parte de la MTV—, alzaran al icónico álbum al número dos de las listas estadounidenses.
Este fue el último momento en que Pearl Jam no fueron nadie. Tocando en su ciudad, para una mezcla de fans tempraneros del grupo y paisanos que pasaban por ahí y no tenían nada mejor que hacer.
Documentazo.
Steve Wooledge retweeted
An analyst who understand their data sources can correct a wrong answer about cancelled orders from an AI query on their enterprise data. But the correction dies with the session, and next week someone else repeats it.
So memory is an entity in Collate with owners, reviewers, certification, and RBAC.
Read more from @CollateData and @open-metadata Co-Founder and CTO, @d3fmacro --> buff.ly/7j0foeS
#Ontology #AIAgents #OpenMetadata
Steve Wooledge retweeted
We improved Claude's accuracy 7x on messy enterprise data. Without changing the model or tuning the prompt.
Using the Spider 2.0 benchmark, built on real production warehouses: messy table names, no documentation, a decade of naming drift. Claude Sonnet scored 10.8% accuracy.
We ran the identical questions again, but with a context layer in place. Accuracy jumped to 76.5%.
A context layer is where you write down what your data actually means, in a form the agent can read before it writes SQL. Without it, the agent sees tables and columns and has to guess. With a context graph, it actually understands your data.
Watch🎥here👉 buff.ly/1eI0lJH
#dataengineering #contextgraph #contextlayer #aicontext
Steve Wooledge retweeted
The merger of @fivetran and @getdbt sent some shockwaves through the data community, but lucky for you, Collate supports both as part of your #datagovernance ecosystem. In this new Collate Solutions, Aydin Geeringh, and @progrockrec will show you what Fivetran looks like in Collate, how to set it up, and how to take advantage of the advanced tools available, such as full #datalineage. This video is not to be missed!
🎥Watch 👉here: buff.ly/soouuAX
#Fivetran #DataCatalog #OpenMetadata #DataIntegration #Collate
Steve Wooledge retweeted
Save the date: July 22 at 9 AM PT for our next OpenMetadata Community Meeting.
Prasanth Cherukuri from @salesforce joins, plus get VIP seating for the next in OpenMetadata and Collate!
Register: buff.ly/u8zrA5O
#OpenMetadata #Collate #AI #DataEngineering
I really enjoyed the conversation around governance in an AI-first world with Lukas from Airbus
Agents don't make bad data better. They make bad decisions faster.
Lukas Patzke of @Airbus on data products and governing for agents as consumers of trusted data. #CollateSummit '26
Full panel discussion: buff.ly/X5wLsbk
#DataGovernance #DataProducts #AIAgents
Steve Wooledge retweeted
30 years of hereditary cancer data. One governance platform.
Dan Kostecki of @AmbryGenetics on governed AI on the most sensitive data there is: a PHI-free environment, 11,000+ assets, fully de-identified under CAP/CLIA/HIPAA/FDA. #CollateSummit '26
Full talk → buff.ly/vHz9Npp
#ClinicalData #DataGovernance #DataProducts
Steve Wooledge retweeted
🚀 Connecting Collate to @StarRocksLabs just got easier. In this episode of Collate Solutions, @progrockrec teamed up with Chris Sapinksi of @phoenixdataai to build a seamless integration that lets you catalog and govern your StarRocks data warehouse in minutes.
You'll learn how quick it is to connect StarRocks to Collage, and you'll also get an example of exploring the ingested assets using our conversational, Ask Collate, AI interface. No more digging through settings menus or wondering how to get back to that spot you saw one time. Just Ask Collate!
🎥👉Watch here: buff.ly/I8D5P6s
#StarRocks #DataCatalog #OpenMetadata #DataIntegration #Collate
Steve Wooledge retweeted
"Blast radius becomes a query, not a search."
Jeppe Johansen of @Unity on Dax, the Slack bot built on the @open_metadata MCP: 7-minute compliance reviews and incident triage in Slack. #CollateSummit '26
🎥 full talk
buff.ly/RdIBdrb
#MCP #AIAgents #OpenMetadata
Steve Wooledge retweeted
"The model isn't the answer. Context is."
Inside Kepler, @OpenAI internal AI data agent: 600+ PB/day, 70K datasets; any employee can ask a question and get an answer in minutes. Bonnie Xu at #CollateSummit '26.
buff.ly/jzfLlhq
#AIforData #ContextLayer #AIAgents
Steve Wooledge retweeted
@OpenAI built Kepler, its internal AI data agent, on OpenMetadata. 70K datasets, 600+ PB/day, question to answer in minutes.
Bonnie Xu on the six layers of context behind it. #CollateSummit '26.
youtu.be/-6czrxuHJLg
#OpenMetadata #DataEngineering #AIAgents
