AI workflows, LLM orchestration, and decision-support systems built for real work. I design tools that help people research, think, and act — not just interfaces that gesture at intelligence. This site is a record of what I’ve built, tested, broken, and learned.
Most AI strategy work happens too far upstream from the system. The credible work starts when someone is actually building, testing, breaking, and improving workflows against real constraints and real use.
Six things I built and use. Not concepts, not slideware — live code with real output, real workflows, and one very demanding user: me.
Watches the accounts I care about for the changes that create a reason to reach out — new hires, role backfills, launches, news that implies a budget or a pain. Scores each signal and queues the few worth acting on as a ranked morning briefing.
LIVEA multi-source intelligence system that watches the AI industry around the clock — funding rounds, launches, partnerships, and talent movement. It turns noisy signals into concise briefs that read like analysis, not a feed dump.
LIVEAggregates analyst reports, vendor announcements, and market signals across the managed-services landscape. Replaces the “read fifty newsletters a day” tax with a single brief shaped by a domain ontology built by hand.
LIVEOne dashboard for revenue goals, journal entries, concept docs, daily actions, and health metrics. Rebuilt twice — once after a full filesystem wipe.
● in useA consumer experiment built from years of study in symbolic systems and pattern interpretation. Still live, but distinct from the AI workflow systems that are the main focus today.
LIVETracks job posts, ad activity, and social signals to identify brands likely investing in UGC. Surfaces likely decision-makers and gives outreach a stronger starting point.
LIVELong-form pieces written while building, not after the fact. Updated when something is worth saying — never on a content calendar.
Most outreach fails because it starts with a list. I rebuilt mine to start with a signal — and turned prospecting into a morning briefing.
Enterprise adoption of agentic AI demands integration depth, security architecture, and respect for the workflows already in place. Capability isn't the gap — fit is.
USB standardized hardware. MCP is doing the same for the wires between agents and the systems they need to act on. The implications compound quickly.
A small log I keep open while building. One-liners, ratios, things I want to remember for later.
The hard part of agent design isn’t the agent. It’s the boundary.
If your eval set fits on one screen, you don’t have an eval set.
Most “AI strategy” work gets sharper after a single shipped prototype.
Latency is a feature. So is silence.
Twenty-five years across enterprise tech. Today I build with AI every day — testing, breaking, and rebuilding the kinds of systems most people are still writing about.
By day, I work in cloud and infrastructure services for financial institutions. This site is where I share my own systems, experiments, and point of view.