Product owner and engineer who turns ambiguous requirements into shipped software. I build AI into every stage of the product lifecycle — living in Claude & Claude Code and defaulting to "can AI do this part?" first.
I run delivery like a product owner and build like an engineer — chasing decisions to closure, flagging risk before it becomes a fire, and reaching for AI on every task that shouldn't need a human.
I don't bolt AI on at the end — it runs through the whole lifecycle, from discovery to shipped code to the retro. Here's where it earns its keep.
Synthesize regulator docs, competitor scans, and user interviews into crisp problem statements — cutting prospect research effort 70%.
Turn requirements into PRDs with owners and acceptance criteria — AI drafts, I edit for judgment. Saves ~8 team hours every sprint.
Prototype and ship real React/TypeScript with Claude Code as a pair — shipped 12+ features on schedule and cut cycle time 25%.
Run AI evals on model outputs and query 50K+ records in SQL to close the loop — resolving issues ~20% faster.
From regulated SaaS to 0-to-1 AI platforms and a bootstrapped D2C brand — here's a slice.
Defined and built an AI platform automating prospect research, outreach, and campaign execution — from CSV lead import to review-ready personalized campaigns.
Built a real-time what-if staffing model across 10+ inputs, scaling 10–1,000 cars over 2 hub formats, with a one-page PDF export ops leads present to stakeholders.
Ran spec-to-ship delivery across eng, design & QA — converting EY and Kuwait CMA requirements into clear specs, owners, and acceptance criteria. Automated status tracking, spec drafts, and reports with Claude.
Ran a 3-SKU D2C brand end-to-end — sourcing, branding, paid ads, fulfillment, CX. Built COD verification and WhatsApp nurture flows that moved the core metrics.
Living in Claude & Claude Code to automate process overhead, draft specs, run evals, and keep delivery moving.