My focus is on the parts of the AI stack that organizations consistently skip — and then regret. Not the model selection, not the prompt engineering. The governance layer: how memory is typed, versioned, retained, and made auditable when agents are running in production against real data.
What I do
I build agentic AI curriculum for technical field teams, design partner enablement programs for enterprise AI ecosystems, and write about what it actually takes to ship governed agentic systems — not just demo them. The AgentOps series is the long-form version of that thinking: a practical guide to production agent stacks, with governed memory as the through-line.
The ARES project
ARES is a typed-memory governance framework — the real architecture behind the audit log on the homepage. It treats memory not as a convenience feature but as an infrastructure concern: every write is typed, every retention decision is attested, every redaction is recorded. The premise is simple: you can't audit what you can't trace, and most agent stacks have no trace.
Background
Twenty years in data and AI — starting from enterprise BI architecture and moving into AI/ML platform enablement. I've delivered at scale for clients including Stanley Tool (a 1,500-user SAP BI 4.0 rollout), Major League Baseball, and Shire Pharmaceuticals. I'm currently Global Technical Enablement Data Scientist Lead at DataRobot, building curriculum and field capability for teams working with production AI systems.
I've worked at every layer of the stack: systems architect, solutions engineer, data scientist, enablement lead. That cross-layer experience is what keeps me oriented toward governance — practitioners know where the gaps are, and the biggest gap in most agent stacks right now is the memory layer.
Get in touch
Use the button above, or send me an email. If you're building agentic systems and thinking about the governance layer, or designing enablement programs for technical teams working with AI tooling, I'd like to hear from you.