7,900+ users · Live on iOS & Android · Built and operated solo
Full-stack engineer shipping production systems end-to-end: architecture, backend, mobile, and AI. I use a collaborative AI engineering workflow to move at small-team speed as one person.
🟢 Open to: Founding Engineer roles · Senior IC (Backend / Systems / AI) · Forward Deployed Engineer · Contract work
A production dating and social platform for sapphic women. Live on both app stores in 16 languages, grown organically to 7,900+ users. Real-time chat plus voice and video calling over Cloudflare Durable Objects, PostGIS location search, a Hosts program where venues and organizers post their own events, and a self-hosted trust & safety pipeline (open source distilled models on my own GPU rig moderate chat, profiles, and photo likeness so user data never goes to a big AI company). ~250,000 lines of Dart. Designed, built, and operated solo.
Tech Stack: Flutter, Cloudflare Workers, Hono, Durable Objects, Neon (PostgreSQL/PostGIS), Cloudflare R2, TensorFlow Lite, ROCm · Status: Live in App Store and Google Play (rebuilt backend relaunched January 2026)
Email and text marketing SaaS for small businesses, priced by sends instead of contacts so small lists stay cheap.
Tech Stack: TypeScript, Stripe · Status: Pre-launch, billing in Stripe sandbox
A local-first AI engineering platform. Dual-agent system (lead developer agent plus a PM sentinel that audits its work), a four-pass planning pipeline, milestone execution matrix, and a tried-solutions journal so agents stop repeating failed fixes. Runs on local models through Ollama.
Tech Stack: TypeScript, Node.js, Ollama, Vitest (76 tests across 17 suites) · License: MIT
A local-first AI companion app. Every conversation runs entirely on-device through a fine-tuned Gemma 4 model via llama.cpp/Metal, so no data ever leaves the phone.
Tech Stack: Flutter, GGUF/llama.cpp, custom LLM fine-tuning pipeline · Status: iOS-first, active development
An e-commerce and logistics platform for a sustainable family orchard, covering the full lifecycle from inventory to fulfillment.
Tech Stack: React, TypeScript, Tailwind CSS, Supabase, Stripe API · Key Features: Secure payment processing, admin dashboard for order tracking, dynamic seasonal inventory
| Domain | Technologies |
|---|---|
| Mobile & Frontend |     |
| Edge & Backend |      |
| AI / ML |       |
| Data & Storage |     |
I partner with AI agents to architect and execute complex systems. I own the system design, data modeling, and infrastructure decisions; the agents handle scaffolding and execution at speed. My agents get direct access to production logs and crash reporting, trace issues, and propose a plan that I approve or push back on before anything gets written. That's how one engineer ships and operates multiple production products at once. See WHY-THIS-STACK.md for the reasoning behind each project's tech choices.
Solo-built doesn't mean solo-process. Every non-trivial change goes through the same discipline a team would enforce: PR review before merge, documented architecture decisions before implementation, and staged rollouts behind feature flags so nothing risky ships straight to production users.
I also run my own inference hardware (Threadripper, 3x Radeon AI PRO R9700, 96GB VRAM on ROCm) for model distillation and private moderation workloads, so sensitive data stays off third-party AI services.
Outside of Utopian Creations, I work in real-time technical operations for Zoox's autonomous vehicle fleet as a Mission Controller: coordinating live incident response across Teleoperations, Fleet Response, and Rider Operations, diagnosing faults, and directing recovery over CLI protocols. That's the muscle for working inside someone else's process under pressure, not just my own.