Selected work.

Architect and technical lead of the distributed AI system that generates and maintains websites at Luxury Presence, a Bessemer-backed Series C proptech company ($74M+ raised) serving 80,000+ real estate professionals. Every generated site has to be indistinguishable from human-designed work — which means the hard problem is making non-deterministic systems produce deterministic-quality output at scale.

• Built the system from the ground up on AI-first principles with a human in the loop. Agents are first-class citizens: autonomous agents manage a site’s whole lifecycle, from creation to ongoing maintenance to quality management.
• Replaced a multi-week manual site-building workflow with an automated pipeline that ships production-ready sites in hours.
• Own the full generation pipeline: prompt design, output evaluation, layout consistency, and brand-fidelity controls.
• Architect of the spec-driven development meta-framework built on Claude Code — treating specifications as the source of truth and using AI agents to execute against them — before the industry converged on the term.
• Lead the engineering team building on the AI foundation and set technical direction for every AI-powered website experience the company ships. Work directly with the C-suite and product leadership on the AI roadmap.

Independent build and consulting practice. Partnered with early-stage founders and SaaS teams to take AI-native products from 0 → 1, shipping across mobile, voice, and developer tooling.

• Designed and launched an AI debugging agent targeting the emerging AI-assisted development segment. Secured 100+ beta signups and converted paid pilot conversations within 8 weeks.
• Partnered with a health-tech founder to build a cross-platform iOS and Android application for postpartum support, integrating AI across the core user experience — from scoping the product with the founder to shipping the AI-powered features.
• Delivered a low-latency voice AI system for a small-business SaaS client, achieving sub-500ms response times and enabling a new real-time customer interaction revenue line.

Led AI engineering at Durable, an AI-powered small business platform serving 100,000+ SMBs. Owned the AI infrastructure, customer-facing AI agents, and engineering team growth — at a moment when the entire AI infrastructure stack was brand new and providers were going down daily from timeouts, rate limits, and outages.

• Designed and built an AI service processing 500B+ tokens per month when the industry had no playbook for operating LLM workloads at that scale. Engineered the retries, fallbacks, queueing, and provider failover that kept it reliable while the underlying APIs were still wobbling.
• Built an internal agent development framework on top of the AI infrastructure — the abstraction layer the rest of the engineering team used to ship customer-facing AI agents without reinventing the plumbing. Everything shipped after that was built on top of it.
• Cut annual infrastructure spend by $375K (75%) through architecture redesign and CDN optimization, freeing budget for product investment.
• Scaled AI services to 100,000+ active small businesses at 99.99% uptime — through the rockiest year of commercial LLM infrastructure.
• Shipped customer-facing AI agents for SEO optimization, content generation, and site improvement — expanding product capability and revenue per account.
• Cut support ticket volume 40% through automated triage and resolution workflows.
• Grew the engineering team from 4 to 11. Built the hiring pipeline, mentorship program, and security practice, including a bug bounty initiative.

Led core services and GenAI R&D at Streamlabs, one of the largest platforms for live streamers. Shipped products used by 5M+ creators, pioneered the company’s first generative-AI products before the current AI wave was mainstream, and drove platform-wide reliability, identity, and security initiatives.

• Promoted from Senior to Principal within two years (2019 → 2021) on technical impact and scope.
• Directed GenAI research and development for Streamlabs and shipped the company’s first generative-AI products — including Logo Maker and AI-powered creator tooling — in the early days of the current AI wave.
• Pioneered new products adopted by 5M+ streamers, including Campaign Manager, Alert Box V2, and a suite of widgets and tools that became core to the creator workflow.
• Led the Streamlabs Identity initiative, unifying authentication across the platform for 5M+ users and eliminating fragmented identity systems.
• Rebuilt the real-time event alert pipeline processing millions of events per day. Reduced P99 backend latency by 65%.
• Replaced the legacy jQuery and iframe widget system with a modern Vue.js architecture, cutting feature development cycle time by 50%.
• Led the core services team responsible for global platform reliability, performance, and availability. Ran the company-wide bug bounty program.