A mix of AI-native products, workflow systems, and complex digital experiences.
A working AI tool that sorts a virtual-care support inbox, gets health messages to nurses in seconds, and drafts replies reps approve. Built end to end and passed all 16 test tickets, backed by 76 automated tests.
An AI agent built on Jev that investigates bank fraud alerts, closes the clear false alarms and sends the rest to an analyst with a recommendation. Tested on 50 held-out alerts: 45 right, no fraud closed alone, at a third of the cost of a single AI model.
An AI-automated system that scripts and produces a complete microlearning course in about 2 hours, with a person approving every script and video. 13 courses at 50–70% lower cost, piloted with 6 clinics.
A practical process for moving from discovery to decisions, build, validation, and launch.
From first prototype to measurable impact, here’s how I help teams ship AI that works.
Agents that investigate, act within set limits and hand the rest to a person.
Tools that sort, route and draft routine work for support and ops teams to review.
End-to-end automations, documented and handed off so teams can own them.
Working AI prototypes that teams can test with real users before engineering commits.
Research that finds where AI earns its place, then scopes and ranks what to build first.
How AI asks, shows how sure it is, recovers from errors and hands off to a person.
Test sets that show where AI is right, where it fails and what each decision costs.
Training, guides and rollout plans that make a new AI tool part of how a team works.
A focused toolkit across product, design, AI, automation, and delivery.
The stack changes with the problem. The goal stays the same: make something useful that works.
Highlights condensed from LinkedIn recommendations by managers, teammates, and collaborators.
With AI as my copilot, I’m no longer limited to screens or imagining what’s possible — I can design, build, test, and bring ideas to life faster than ever before.
