Mapped every ticket category, built a hybrid RAG + classifier pipeline, then integrated directly into their Zendesk stack. The team went from 400 daily tickets to 128 — with CSAT up 14 points.
Built a document intelligence layer that ingests raw portfolio data and outputs analyst-grade prose narratives. Saves 40+ hours per quarter.
Designed and deployed a fine-tuned generation pipeline with brand voice guardrails, reducing copywriting cost by $240K annually.
End-to-end private deployment on NHS-compliant infrastructure. Clinicians reclaim 90 minutes per day on average.
Four engagement types. Clear scope, honest timelines, and a team that can own it when we're done.
A 2-day intensive that maps your highest-leverage AI opportunity, sizes it honestly, and produces a 90-day roadmap your team can execute.
From scoped problem to production system. I embed with your team, build the MVP, and hand over with full documentation and evaluation harness.
End-to-end delivery for complex, multi-system AI initiatives. Typically 3–6 months. Includes MLOps, monitoring, and change management.
Ongoing strategic partnership. Two days a week, embedded with your leadership. For companies serious about making AI a core competency.
I've spent eight years at the intersection of machine learning engineering and business strategy — first building recommendation systems at scale for a Series D SaaS company, then leading AI transformation at a global management consultancy, and now working independently with a small portfolio of high-growth clients.
My approach is direct: I find the highest-leverage AI opportunity in your business, size it honestly, and build the systems to capture it. Engagements end when the work is done and the team can own it.
I publish a weekly newsletter on applied AI strategy read by 8,000+ operators, and I speak at events including SaaStr, Collision, and the FT's Digital Futures series.