Ray Festa

• AI systems consultancy: client-facing intake systems, workflow automation, and oversight layers that decide what an agent finishes alone and what reaches a person.
• Design, build, and deploy end to end, from the interface through the agent architecture to the evaluation loop that grades its output.

• Designed management and oversight surfaces for managed database and data-infrastructure products used by Fortune 500 companies and government agencies.
• Visual data model builder for time-series schemas: 93% adoption at launch, and setup time compressed from roughly 90 days to hours by meeting users inside their existing SQL mental model.
• Conversational analytics tool rebuilt as a progressive dialogue: query accuracy went from 20% to 90% after the redesign, with prompt tuning done alongside the engineering team.
• Moved from configuration consoles to agentic oversight: the surfaces where people supervise AI agents, set thresholds, and review what the system did before accepting it.

• A decade of product design across B2B and consumer products before specializing in enterprise data infrastructure.

Alpine Velo

AI cycling coach, live on the App Store. Designed and built solo: the iOS app, the coaching agent, the training science it reasons with, the brand and visual system, and a self-grading loop that scores the coach's answers and rolls back changes that make it worse.

AI Finance Dashboard

Production-ready UI for AI finance agents. Nine pages and a documented design system, built through Figma, v0, Claude Code, and Vercel.

Capium

Personal finance app, designed but never shipped. Kept as a craft piece: the interface argument for showing a user what is true before telling them what to do.