I Design How Teams Should
Build Software with AI.
AI tools are fast and powerful. Without architecture, they produce expensive chaos. I work as a Lead Architect to help CTOs and VPs of Product ship AI-powered software that actually works — from Idea to Architecture to AI to Launched.
Every AI team hits the same wall
The first phase of AI adoption feels like acceleration. The second phase feels like drag.
At first, the team ships fast. Prototypes appear overnight. Demos impress. Then it stops working. A generated abstraction is locally correct but doesn't fit the system. Different developers produce different patterns because the prompts are different and the architectural constraints — if there are any — are vague. The team isn't the problem. The tools aren't the problem. The missing architecture is the problem.
AI doesn't remove the need for architecture. It raises the cost of not having one.
Architecture before Prompting
The teams seeing real results from AI-assisted development aren't the ones with the best prompts. They're the ones who defined the architecture first. Data model, coding standards, component patterns, API contracts — the foundation every AI tool builds on top of. Without it, AI output becomes a patchwork of conflicting patterns. With it, AI tools produce coherent, production-grade code.
Define the foundation before asking AI tools to generate code.
The architectural decisions are the context for the AI tools
Review the AI output against the architecture
One Engagement. Four Phases.
Idea
Deep dive into your business context, existing systems, and where AI fits. We define what success looks like.
Architecture
Designing the data model, API contracts, coding standards, and component patterns. Every decision documented with its rationale
AI
Building a validated slice of the system using the architecture I designed. Bad assumptions surface here, not six months into production
Launched
Your team gets the architecture, the standards, and the working code. I walk your team through the system, answer their questions, and make sure the handoff is real
Case Studies
Selected Projects
Industrial Scientific
Client - Enterprise
Gas detection and worker safety monitoring platform used by field workers, safety managers, and system administrators across industrial environments. Built three foundational systems the engineering team builds on: Zustand-based global state managing device data, fleet status, and alarms; global error handling with recovery flows; and a UI component library based on Tailwind’s Catalyst template.
Idea Genius
AI Product
AI-powered startup development platform that guides entrepreneurs from initial concept through launch. Multi-step RAG pipeline threads context across 40+ steps, where each step builds on every decision before it. Config-driven AI orchestration with structured outputs and extended thinking for complex business analysis.
Full Answer
AI Product
AI-enhanced search platform with persistent context. Unlike standard AI search that forgets everything between queries, Full Answer maintains the thread of thought between queries, allowing for a more coherent and accurate search experience.
Audacy
Client - Streaming
Streaming audio platform serving millions of listeners. Introduced C4 architecture modeling to an engineering team shipping without shared architectural language. Gave product and engineering a common vocabulary for system design.
Production Genius
AI Product
Film and video production management platform for content teams. Manages the full production lifecycle — scripts, media assets, cast and crew, scheduling, and deliverables. Designed on the shared Next.js/Firebase template architecture that powers the entire product portfolio.
Crew Genius
AI Product
Design-first workforce management SaaS for retail and hospitality, built on a crew-first philosophy that leads with people, availability, and coverage gaps before scheduling shifts. Consolidates crew information into a single person detail page with six tabs, and ships on a shared Next.js template architecture that powers a family of SaaS products.
Who I Work With
This engagement works best when you're a CTO or VP of Product at a 100–500 person software company. Your team has AI tool licenses. You're three to six months in. Velocity hasn't matched the promise.
You're a CTO or VP of Product at a 100–500 person software company.
Your team has shipped AI prototypes that impressed in demos and fell apart in production.
Your engineers are talented and don’t need more developers — they need architectural direction.
Your AI roadmap includes features but nobody has defined how they integrate with your existing system.
Ready to Build Your AI Vision?
A 30-minute conversation will tell us both whether this is the right fit. No pitch — just an honest look at where you are and what would actually help.
Book a Free Discovery Call