Enterprise AI product leader building agent workflows, developer tools, and enterprise integrations.
I’ve worked across product leadership and technical delivery at Oracle and Tealium. These days I’m building TheTechStack and working on public projects for inspecting, testing, and running agent workflows.
I use AI throughout development. I make the product decisions, review the code, and test what it does. You’ll find a mix of working applications, reusable skills, and exercises below.
- Product and deployment decisions: explore TheTechStack and its public data contracts.
- Working code and tests: inspect the context-packet builder and its behavioral tests.
- Full-stack implementation: explore Support Operations Workbench, a ticket application I use for development exercises with AI.
I built and operate TheTechStack, an independent research and planning resource for teams choosing, connecting, and operating enterprise AI tools.
I use agents to turn cited public sources into structured stack records and editorial drafts. Records go through schema validation, and I review the work before publication. The product and orchestration code are private. The public data contract, evidence model, and process for companies to request corrections live in thetechstack-data.
Reusable skills for researching public sources, preparing content for human review, validating structured records, and handing context between agents. They work across providers and include synthetic examples and validation checks. The context-packet builder also has a Python script and tests you can run.
Exercises for working with coding agents: writing repository instructions, gathering context, reviewing proposed changes, delegating work, and checking the results.
A Next.js ticket workbench with typed route handlers, validation, persistence, and domain tests. I use it to work through ticket investigations and scoped code changes with AI assistance. There’s no AI model running inside the app yet.
Some older projects from my time at Tealium, connecting customer-data systems through APIs and working through the details of getting those integrations to work.
- Magento 2 integration
- Clearbit enrichment prototype
- HubSpot and Constant Contact integration
- Gatsby product demo
These are historical prototypes and integrations, not current production packages.
I like starting with a customer problem, building a small version, and seeing where it breaks. For agent workflows, that means checking what context the agent gets, which actions need approval, and how we’ll know whether the result is right. I document those decisions and the limits of what I’ve tested alongside the code.