The pattern underneath the path
I tend to choose environments where no single skill is enough.
Architecture asked me to think in structure, constraints, aesthetics, and human behavior at the same time. Luxury retail trained me to read trust, timing, motivation, and commercial intent in real conversations. CrossFit kept me in a system where progress is measurable, feedback is immediate, and there is always another layer to learn.
That is the throughline: I like systems where judgment has to be trained, not assumed.
Why I went toward people
I grew up in environments where the default path was technical excellence: top schools, top classes, engineering minds, and a strong pull toward large tech companies.
I respected that path, but I knew the skill I needed to train most was different. I wanted to understand people directly: how they decide, how they trust, how they hesitate, and how they reveal what they value.
That is why I chose commercial work. Sales gave me a live environment for observing human behavior under real stakes.
Why CrossFit stuck
CrossFit stayed with me because it combines things that are usually trained separately: strength, endurance, skill, mobility, recovery, and mental pacing.
It is also measurable. Benchmark WODs make progress visible. You learn what changed because you can compare it against something real. The feedback is direct: you either moved better, paced better, recovered better, or you didn't.
That kind of system is hard for me to leave. It gives structure to effort, and it turns consistency into data.
The community matters too. It creates a kind of belonging around shared effort - people doing difficult things together, with no need to over-explain why it matters.
When the gap became too visible to ignore
Inside a high-touch sales department, I kept seeing a workflow gap that repeated across the team.
No one had assigned the problem to me. There was no budget, no roadmap, and no dedicated team. But repeated friction is exactly the kind of signal I pay attention to, so I started mapping the workflow behind it.
I had to understand what infrastructure already existed, where the client data moved, who needed access, what had to be logged, and what security and compliance stakeholders would need to see.
Then I built the replacement: a Chrome extension that now runs in production across the department. I scoped it, designed the workflow, built with Claude Code, tested with users, documented data handling, defined access controls, established audit trails, and took it through IT, data, security, and compliance review.
It replaced a $70K/year SaaS at zero incremental cost.
How I use AI
Since AI became part of my work, I finally felt like there was something that could keep up with the speed and shape of my thinking.
I use AI less like an oracle and more like an extension of my hands: a way to move through a system faster, trace logic, test assumptions, expose weak points, and quickly validate whether an idea can become real.
The Chrome extension was built with Claude Code while I wrote, debugged, and iterated across the codebase. AI helped me stay inside the problem longer instead of constantly stopping at the edge of my own implementation speed.
That pattern also shows up outside work. I connect my health data to the Claude API each morning to interpret training readiness before CrossFit. Across the conversational AI tools I use day to day, I set them up to answer in a framework-first way, because that is how I think: define the structure, name the variables, then reason through the system.
For me, AI is most powerful when it turns scattered signals into a structure you can think with, test against, and improve.
From the Chrome extension to a functional AI Sales Agent prototype to a dual intelligence learning system — built across two months, from zero, through iterative reasoning with AI. The speed at which I can move through an unfamiliar domain has fundamentally changed.
How I work
My default mode is framework-first.
Before I optimize a part of a system, I need to understand the shape of the whole thing: the actors, incentives, constraints, data flows, failure points, and feedback loops.
I am most effective when I can own a problem end-to-end. I like ambiguous problems with real users, visible friction, and enough room to design the operating layer underneath.
When I see repeated inefficiency, I treat it as a design problem. I want to know why the current system produces that friction, where the leverage point is, and what kind of tool, workflow, or structure would make the better behavior easier.
Where this is pointing now
I want to keep building systems for environments where human judgment is valuable but trapped inside inefficient workflows.
The areas I am most drawn to are AI workflow systems, internal tools, client experience, CRM, commercial operations, and high-touch service environments.
I care about problems where the work has to become real: mapped, built, tested, adopted, and improved through feedback.