I bridge operations, technology, and the people who rely on both.
Business Operations · Process Improvement · AI-Enabled Operations
I've always been drawn to work where the process isn't finished yet — where the requirements are unclear, the workaround has become the workflow, and someone just needs to sit with the mess long enough to see the shape of a better system. That's true whether it's a support queue at a Fortune 500 bank or a product I'm building from scratch on nights and weekends.
By day I'm a Senior Specialist, Business Technology Liaison at BNY Wealth Service Delivery, where I design and build support and automation systems on Microsoft's Power Platform. Outside of it, I'm the founder of pagarii and the builder of Bowline — the same pattern each time: walk into ambiguity, find the recurring pain, and build the structure that makes it stop recurring.
This is the pattern that repeats across all three case studies below — not a philosophy, just the sequence the work actually follows.
Start with what's really happening — user surveys, review analysis, watching the workaround in the wild.
Separate the one-off complaints from the failure mode that keeps coming back.
Turn the pain into written requirements, user stories, and data structures before building anything.
Build the smallest working version — a Power App, a core loop, an onboarding flow.
Test against real use — release-night scenarios, daily dogfooding, cross-model review gates.
Write the system down so it survives handoff — knowledge bases, versioned specs, governance docs.
Move the repeatable parts to flows, agents, and pipelines — routing, notifications, builds.
Instrument the result so the system learns from real data instead of guessing.
An end-to-end intake, routing, and knowledge system for a 1,200-person internal user base, built on Microsoft's Power Platform — real, shipped work at BNY, recreated here with fictional data. Includes three working interactive mockups: the ticket app, the SharePoint list, and the Power Automate flow.
Read the case study → Founder-Led Product Development · 2014–2026A verified, video-first dating platform built on one honest introduction a day — 378 versioned spec documents as source of truth, a Claude + OpenAI cross-model review pipeline, and parallel AI agents merged back into canon under her direction. In active pre-launch development.
Read the case study → Product Case Study · 2026A focus-and-follow-through system for a genuinely full life, designed and built end to end — web app, native iOS, and an AI planning engine governed by one rule: propose, never auto-add. A private, single-user product in daily use.
Read the case study →Business Operations · Process Improvement · Change Management · Program & Product Operations · Business Analysis · Requirements Gathering · Agile User Stories · Workflow Design · Knowledge Management
Power Apps · Power Automate · SharePoint Lists · Microsoft Forms · Microsoft Access · UiPath · Microsoft Copilot · AI Builder · OneNote
Claude Code · Claude Cowork · ChatGPT · Gemini — requirements development, process design, and operational analysis, with independent validation of every output
Excel · Salesforce · Jira · SPSS · Stata · Python & SQL (working knowledge, AI-accelerated) · cluster analysis · regression · topic modeling · survey design · qualitative interviews