Explainable product matching: a 0–100% fit score without a black box
Why a deterministic scoring engine beats an opaque model for trust, debugging, and "why this fits you" reasons.
Insights
Notes from the build.
Practical writing on AI in products, platform decisions across commerce, SaaS, and mobile, and how we ship faster with AI in the loop. No hype — just what we’ve learned shipping it.
AI Engineering
AI Engineering
On-device AI: AR try-on and fit models that never leak body data
Running biometric, voice, and body models on the phone so personal data stays put — privacy as a feature, not a disclaimer.
AI Engineering
Rules-first, LLM-second: AI-assisted field mapping for messy data
A pragmatic pattern for enterprise data integration — deterministic suggestions first, an LLM fallback, audited end to end.
Platforms
Multi-tenant from day one: isolating data while sharing one platform
Per-tenant databases, entitlements, and the architecture trade-offs we'd make again.
Platforms
MACH & composable: when headless pays off — and when it doesn't
A decision framework for choosing composable over a monolith for D2C and marketplaces, or staying put.
Product
Shipping a ride-hailing MVP: rider, driver, and the trip between
What it actually takes to get a two-sided mobility app — booking, live trip, and fare — to a real first ride.
Delivery
How AI in our toolchain cut a build cycle by 40%
Where AI-assisted discovery, code, and QA actually moved our delivery timeline.
Subscribe
Get new field notes in your inbox.
Occasional, practical writing on AI-first product engineering. No spam — unsubscribe anytime.
Ready when you are
Prefer to talk it through?
Book a call and we'll dig into the specifics of your AI-first roadmap.