I design the systems that act on what the model decides.
That means the interface, the logic, the escalation paths, and the feedback loops that determine whether AI behavior can be seen, questioned, and stopped.
I work across product, engineering, data, policy, and executive decision-making to design systems where accountability is traceable, decisions are defensible, and humans stay in meaningful control - especially in regulated or high-stakes environments.
Thank you for spending time with the work.
Wells Fargo
Designed decision-governance workflows that made probabilistic model behavior reviewable, contestable, and accountable inside a high-stakes underwriting environment.
Automation Anywhere
Led product and systems design for automation workflows where visibility, trust, and operational control determined whether human-in-the-loop AI could be used safely.
Netscape / AOL
Early work across personalization, messaging convergence, and large-scale onboarding helped make the internet feel usable, familiar, and trustworthy for ordinary people.
Friendster
Inside the first real social network, unstable infrastructure, shifting product identity, and early social graph behavior revealed how quickly users could define a platform faster than the company itself.
Wells Fargo
Inside Wells Fargo AI Enterprise Solutions, the work focused on operationalizing trust in regulated AI systems through governance frameworks, human oversight, interpretability, and scalable experimentation across 17 business units.
A real-world case study from Autodesk showing how fragmented legacy systems were audited, aligned, and unified into a cloud licensing platform through influence, transparency, and system-level design—an approach that continues to hold across modern enterprise and AI systems.
AI systems rarely fail in obvious ways. This paper defines a simple operating model for structuring signal, interpretation, escalation, and response so decisions hold under real-world conditions.
As AI collapses the distance between design and production, the handoff disappears. Design no longer proposes. It decides, making accountability and system behavior inseparable from what ships.
Software has moved from recording and recommending into deciding, but most systems were never designed to handle the responsibility that shift requires.
A model can score risk in milliseconds. Deciding what to do with that score is where governance begins.
Inside a Santa Clara County vote center, elections operate as human-governed decision systems that absorb uncertainty, preserve participation, and produce auditable outcomes.
Exit Wounds extends the themes explored throughout Amid the Noise into poetry: recovery, toxic attachment, substance use, homelessness, and the cost of naming what survival required.