Events and Webinars
Building a Context Layer for AI Agents
8.22.26
On this episode of DM Radio, host Eric Kavanagh talks with Matterbeam CEO Michael Kowalchik about why AI agents keep hitting old data problems. A stale table or an ambiguous definition can quietly turn a capable agent into a confident source of bad decisions, and those errors compound across multi-step workflows. They discuss how an immutable, replayable fact log and agent-specific governance turn promising prototypes into systems you can trust in production.
Responsible Use of PyTest
1.21.26
At the January Boston Python Meetup, David Sturgis, Senior Engineer at Matterbeam, explores PyTest—the modern Python testing framework that's more Pythonic and flexible than traditional unittest. Learn how PyTest's test discovery works, master fixtures and parameterization, and discover strategies for organizing tests effectively.
View the slides only
From Data Hoarding to AI-Ready
1.8.26
This webinar is perfect for data leaders and AI/ML teams facing massive storage costs for unused data, long wait times when AI initiatives need data access, constant pipeline rebuilding as requirements change, or struggles preparing clean training datasets for AI projects.
Data Migration Without the Risk
12.4.25
This webinar is perfect for data and engineering leaders dealing with vendor lock-in or spiraling costs, post-acquisition system consolidation, legacy modernization, or AI data prep.
Why Your Data Is Failing You
12.2.25
Michael Kowalchik joins the Between Product and Partnerships podcast to talk about why traditional data infrastructure breaks down once a business hits real complexity. The conversation covers immutable logs and replayable streams as an alternative, and why the current wave of AI adoption makes it critical to trace what happened, reproduce it, and understand how data moves through each step.
Data Products, On-Demand
9.12.25
Matterbeam CEO Michael Kowalchik joins DM Radio alongside Mark Madsen of Third Nature and Josh Pendergrass, VP of Engineering at Broadlume, to talk about building data products on demand. The discussion covers how capturing enterprise data from any database in an append-only log lets teams stand up focused data products quickly, without the overhead of another warehouse or repository.
Stop Fighting Your Data Infrastructure
Most companies we talk to know something’s broken. They just assume six-month timelines and ballooning costs are normal. They’re not. Let’s talk about what’s actually possible.
