journey
what i've built — ai systems at scale, and the teams and businesses around them.
work
ai systems at scale — built, led, grown
graph-based ML over transactional data — 20M+ entities, 25M+ relationships, 50M+ daily users
data-intelligence layer behind Manchester City fan onboarding — caught 10% flow defects, 25% drop-off across a 100M+ fanbase
founded an autonomous agentic orchestrator — small LMs, spec-driven, red/blue-team agents
shipped @bosun-sh/logbook open source — fs-based kanban for AI agents (MCP + CLI); published 2 preprints
founded an AI studio — 8+ systems for an ARG agency and 4 startups, prototypes in <3 weeks; 2 clients hit funding
growth engineering: event instrumentation + AI querying layer → 100× activation, churn under 5%
stood up a startup's AI cell in <3 months — stack, standards, hired 5 engineers
led a 15-person cross-functional team — RAG lesson planner, tutoring chatbot, vector content matcher to prod
−50% teacher churn, activation 50 → 1,000+
OpenAI-powered CX chatbot a year before ChatGPT — −25% agent workload
confidence-scoring + context-injection layer before orchestration tooling existed
fine-tuned T5 for domain QA on EC2; GPT-2 few-shot classifier — in-context learning before it had a name
enterprise scale — Live Nation events at 1M+ concurrent users, global e-commerce migration, incident recovery, +15% team capacity
papers
published research on ai engineering and governance