Built to handle the world's most critical data, Redpanda now governs the AI agents that act on it.
Global
Joined October 2021
- Tweets2.6K
- Following183
- Followers5.3K
- Likes708
Pinned Tweet
You can't govern an AI agent with a prompt it's allowed to ignore.
Most governance stops at the application boundary. Agents don't. Every handoff is a governance gap.
The Redpanda Agentic Data Plane closes them: one governed layer, every request.
See it in action ➡️ bit.ly/4zKzik0
Your strategy is only as fast as your slowest data pipeline.⚡
In quant trading, every millisecond of delay and every dropped message shows up as slippage, missed fills, or decaying alpha. AI now raises the stakes further: models retrain on live market data, and agents are moving into research, risk, and surveillance.
Our new handbook shows how @NYSE, @jumptrading, @AlpacaHQ, @StoneX_Official and @zafin run real-time streaming and agentic AI in production.
Download the handbook 👉 bit.ly/4yzcF0a
Redpanda’s Josh Purcell joins Daniel Davis and Dustin Gouker at NEXTPredict NYC to tackle the question: predictable latency, massive fan-out, and clean recovery under real load are the key. 📈
🗓️ Thu, Oct 22 · 12:30 PM ET
📍 Hudson Yards, NYC
Come say hi 👋 bit.ly/4hZxQ6I
ALT Promotional graphic for NEXTPredict panel discussion event with the title, "How do you upgrade a market that can't stop trading?" The event is scheduled for Thursday, October 22, at 12:30 PM, located at Hudson Yards, NYC.
Your data has a shadow. And that’s a good thing. 🕶️
Next week on Data & AI Weekly, Redpanda’s Peter Corless and Trevor Blackwood dig into Shadowing. Join live at 9 a.m. PT to see real-time data replication keeps production running through disaster recovery, migrations, and more. Plus a look at how leading enterprises are already using it.
Sign up once, catch every episode 👉 bit.ly/4qoGyOp
ALT Promotional banner for "Data & AI Weekly" featuring speakers Trevor Blackford, a Senior Product Manager at Redpanda, and Peter Corless, a Principal Product Marketing Manager at Redpanda. The topic is "Shadowing: How real-time data replication supports operational resilience." The session is on October 27 at 9 AM PT with a "Register Now" button.
Redpanda retweeted
We just shipped @redpandadata support for oleander 🌸
You can now stream Redpanda topics into your oleander Iceberg catalog using a Redpanda Connect pipeline.
With this integration, you get:
→ Streaming data. Query-ready in Iceberg
→ One stream. Multiple open compute engines.
→ Secure agent access to streaming data via MCP.
Learn more on the integration and how to get started 👉 docs.oleander.dev/integratio…
#DataCompute #ComputeRouting #MultiEngine #MultiEngineDataWarehouse #ApacheIceberg #Redpanda
See what’s next. 👀
The Agentic Data Summit agenda just dropped: streaming at scale, agent guardrails, Kafka-to-Redpanda stories, the roadmap, and a live AMA with leadership.
Free to attend. Stream live from wherever you are → bit.ly/4h9FDhG
CoreBreak proved model-level guardrails fail the moment you skip the model.
Next week, Redpanda's Tyler Akidau and Marc Millstone break down the attack, the fix and where guardrails really belong.
📅 Tuesday, October 20 at 9:30 a.m. PT
👉 Join us live: bit.ly/4qoGyOp
ALT Announcement for 'Data & AI Weekly' event on October 20 at 9:30 AM PT, featuring CoreBreak topic. Includes speakers Tyler Akidau, Chief Technology Officer of Redpanda, and Marc Millstone, Head of AI at Redpanda. Includes a 'Register Now' button.
Would you hand a brand-new hire root access and trust them to enforce your security policies on day one?
That’s what we’re doing when we trust AI agents to carry access rules and data classifications themselves.
@AIconference San Francisco, Redpanda VP of Engineering Piyush Agarwal explained why that’s risky and showed a better approach. Redpanda’s Agentic Data Plane uses out-of-band metadata channels to enforce governance entirely outside the agent’s read/write path. It scopes data access going in, constrains actions during execution and captures tamper-proof audit trails coming out.
Watch the full talk: bit.ly/4ymjGkX
ALT Graphic for The AI Conference featuring Redpanda. Text reads: 'Watch the replay: Introducing Redpanda’s Agentic Data Plane.' Includes a photo of Piyush Agarwal, VP, Engineering at Redpanda, with a button labeled 'Watch Now.'
Tomorrow, someone might ask who owned the decision your AI agent just made.
Will you have an answer? 🤔
Redpanda's Melissa Czapiga and Daniel Takabayashi go live tomorrow October 13 at 9 a.m. PT on Data & AI Weekly to cover what you need in place before an auditor, customer or board member asks.
🔗 Grab your spot: bit.ly/4qoGyOp
ALT An announcement for "Data & AI Weekly" at 9 AM PT on October 13, 2026. Topic: "When AI Governance Fails: Who’s Accountable When Agents Get It Wrong." Features Melissa Czapiga, Chief Marketing Officer, and Daniel Takabayashi, Staff Forward Deployed Engineer, both from Redpanda. Register now button included.
Software used to make you adapt to it. Now Agents adapt to you.
@usekernel CEO Catherine Jue on why 99% of the web wasn’t built for agents, and why live context changes everything.
🎧 The latest episode of "Hello, Agent!" is now live: bit.ly/4yyYlpa
Manufacturing can't pause for a migration. These 4 didn't have to:
@GlobalFoundries: decisions in seconds, not hours
@RangeEnergy: went live in 30 min
Rivian and Volkswagen Group Technologies: escaped a cluster-killing Kafka bug
@johnsoncontrols: $1.6M saved/yr
Learn more: bit.ly/4zGn7nJ
Gartner expects weak accountability, not model performance, to be the top cause of AI failure through 2028. Is your organization doing what it can to prevent it? 👀
On the next episode of Data & AI Weekly, Melissa Czapiga & Daniel Takabayashi cover what your team should have in place before someone asks who's accountable.
Set your reminder 🔔 bit.ly/4i6yZZZ
A lot of teams run streaming in one system and analytics in another, then add AI agents wherever they fit. It works, but every handoff adds latency, cost, and one more thing to manage.
In December, we're spending a day showing how running all three on one platform can help your team work smarter. Hear from the engineers, architects, and Redpanda customers who are seeing the difference.
Grab a free spot and join from anywhere → bit.ly/4h9FDhG
This made our whole week 🫶
If you've been meaning to try Redpanda, here's your sign: Kafka-compatible streaming built for real-time data, and now powering AI agents.
Get started in minutes and tell us what you build 👇
Last chance! 🗣️ Data and AI Weekly goes live tomorrow at 9 a.m. PT.
Founding members of the new Streamhouse Working Group join Redpanda's Peter Corless to break down what a streamhouse architecture needs.
🛟 Save your spot: bit.ly/4qoGyOp
AI agents escaped a sandbox and hacked Hugging Face. Nobody noticed until Hugging Face said so.
A kill switch on the model wouldn't have helped. Your agent boundary has to live outside the model, and you have to own it.
See why 👉 bit.ly/4ARS1ea
Every new hire gets rules. Your AI agents should too.
Learn how to move agents from pilot to production 👉 bit.ly/4AHcEtq
Your AI agent is only as good as the data it can see.
Redpanda SQL is Postgres-compatible and available now on Redpanda BYOC.
See what it solves 👉 bit.ly/3V7n0m4
Lakehouses are great at telling you what happened. They're not built for what's happening right now, and that's exactly the gap AI agents are exposing.
Join Redpanda's Peter Corless as he moderates a panel with Redpanda's Matthew Schumpert, @streamnativeio's Kundan Vyas, and @VervericaData's Fabian Wilckjens on next week's episode of Data & AI Weekly.
They're digging into what it actually takes to build a Streamhouse architecture, and what's next for the Streamhouse Working Group.
📺 Tune in live on Tuesday, September 29 → bit.ly/4qoGyOp
What a way to kick off @BigData_LDN! 👏
Thank you to everyone who joined us earlier today to hear from Redpanda's own Josh Blakeley during his session, "Governing AI Agents in Practice."
There is still time to connect with our experts. Drop by Stand L108 throughout day two of the show to learn more about how we're helping enterprises govern agents in production and giving teams predictable, scalable governance at scale.
