I advise like someone who has to build it.

Founded Crosby33 as the operating home for AI systems consulting, products, and open-source work. I advise founders and teams on what to build, design the architecture that will hold, and remain hands-on through production delivery.

• • Lead strategic assessments and architecture decisions for new and stalled AI systems, including the explicit recommendation not to build when that is the sound business decision.
• • Work as an embedded fractional CTO and forward-deployed engineer across roadmap, architecture and tooling, developer enablement, and output validation.
• • Build and publish Crosby33 products, experiments, open source, and reusable agent skills as direct evidence of the practice.

Full-stack engineer at a digital mental health platform serving 40+ organizations and 300,000+ covered lives worldwide. Primary resource for agentic development practice, cloud backend architecture, and AI systems across the engineering team — pioneering the shift from AI as a side tool to AI as core development methodology.

• • Championed Claude Code adoption from personal tool to team-wide standard, then introduced CodeRabbit to automate PR reviews — resulting in 63% PR throughput increase and eliminating manual review bottlenecks for staging deployments.
• • Planned and executed a zero-downtime migration of 200+ Firebase Cloud Functions, orchestrating up to 4 parallel Claude Code instances via git worktrees. Established Infrastructure-as-Code with Terraform, cutting CI/CD deployment time by 50%.
• • Executed a 9-dimension platform modernization entirely through Claude Code — CRA to Vite, React 17 to 19, MUI v4/v5 to v7, and 20+ major dependency upgrades across 1,293 files (~92K net lines removed). 100% AI-generated initial pass with only ~20–25 regressions in QA.
• • Built an AI investigation agent using Mastra orchestration with two custom MCP servers (BigQuery + Firestore) that reduced video call issue diagnosis from 30–60 minutes of manual investigation to under 5 seconds of automated analysis.

Led engineering for an AI-powered meeting intelligence platform, driving both technical modernization and the integration of generative AI capabilities. Built one of the earliest production Claude integrations in spring 2023 — before tool use or JSON mode existed, when reliable structured output meant creative prompt engineering.

• • Built an automated 4-tier meeting summarization pipeline and conversational Q&A assistant using Claude and AssemblyAI's LemUR API with parallel Cloud Pub/Sub orchestration — among the first production GenAI systems of its kind.
• • Re-engineered the backend from JavaScript to TypeScript with event-based microservices, achieving over 90% crash reduction. Modernized the React front-end with a custom design system and migrated to trunk-based branching for continuous delivery.

Led a global engineering team of 13 building a preventive healthcare platform on GCP. Responsible for backend architecture, infrastructure automation, and scaling systems to handle high-volume clinical operations.

• • Engineered a composable finite state machine service with GraphQL interface integrated with idempotent microservices and large-scale message queue — the core orchestration layer for clinical workflows.
• • Automated GCP infrastructure deployment with Terraform for swift multi-region replication. Migrated a critical registration API from Python to Go to handle high-volume sign-ups reliably.

Co-founded and built a cloud-native financial planning SaaS from the ground up. Owned the full technical vision — architecture, development, hiring, and client-facing delivery.

• • Designed and built the entire Firebase/Cloud Functions backend, delivering a functional MVP in under 3 months that secured the company's first customer contract.
• • Recruited and led a freelance engineering team while partnering on product roadmap, client presentations, and pre-sales activities.

A chapter that shapes how I work with business stakeholders today. Held executive roles across hotel operations — from managing P&L and business strategy for a 3,500-key hotel complex to leading a financial turnaround as General Manager. This is where I learned to translate between operational reality and technical possibility.

• • Led a GM-driven financial turnaround that delivered 76% revenue increase in year one, combining operational restructuring with data-driven decision-making.
• • Managed cross-functional teams of up to 70 and held full P&L responsibility — the kind of business-side experience that makes working with enterprise stakeholders feel natural, not learned.

Managed operational and business strategy for the North American Equity Derivatives business on Wall Street. This role built the instinct for understanding how technology decisions affect business outcomes at scale — a foundation that still informs how I approach engineering problems.

• • Oversaw $300M+ in annual revenue across two trading desks, managing the intersection of technology, operations, and business strategy.
• • Drove $11M in cost savings through vendor negotiations and operational optimization across the derivatives business.