Learning design · AI enablement · Product · Queens, NY · English & Spanish
I help organizations decide where AI is worth the investment, get people using it well, and measure whether it actually worked. I come from educational psychology and learning design. Over sixteen years I've led discovery for new products, launched a platform to 2,000 teachers, designed and run learning for professionals, and evaluated what actually changed. Now I do that work with AI.
AI upskilling for a research lab · 2026 · work samples → I scoped it on a staff survey, piloted the workflow with five colleagues on their real work, and turned what they ran into into the design of a hands-on workshop for about 30 staff, where six teams built five prototype concepts. Since then I've supported the lab's AI forum.
Designing an AI coach · 2025–2026 · design record → Reviews of an AI prototype kept turning into arguments about preference. I mapped the experience the AI would sit inside, classified every touchpoint as human-essential or an AI opportunity, and wrote the behavior specification the team judged the prototype against.
An AI product for entrepreneurs · 2025–2026 · working documents → · how I ran it → I took over an early-stage product partway through, made user research the first priority, and tested positioning with a landing-page test whose thresholds were written before it ran. Volume missed while quality held, so I recommended a pivot to community-first and ran a twelve-week pilot with founders.
Nursing curriculum diagnosis · 2021–2022 For a university nursing school, I tagged around 500 learning objectives against national competency standards, then designed and ran a six-part virtual workshop series with faculty and leadership over seven months. The checklists went to all faculty.
Live professional training at an accounting firm · 2020–2021 I ran the operations of the firm's live and virtual training, about 20 learners a class and at least three classes a week, so the subject-matter experts could focus on teaching. I coached them on explaining complex topics and tracked their license-renewal credits; no one missed a renewal.
Launching an assessment platform, Innova Schools · 2017–2018 Leadership wanted dashboards; research in schools showed the data behind them was entered late and wrong. I made accurate data the first milestone, piloted a mobile tool in five schools, and launched to all 20+ schools and 2,000 teachers a month later, with the pilot IT supervisors as champions. Input errors fell 30% after the first iteration.
Evaluating learning platforms, Innova Schools · 2016–2017 I designed the controlled experiments leadership used to decide which platforms to scale, measuring learning with standardized testing alongside adoption, because usage is not comprehension.
roomread: a workshop tool. Put researched options in front of a room, have each person back a few and name the assumption that could kill one, then read it back live. One page and a Sheet.
Outside my day job I build small tools with Claude Code, to learn how to build product with AI by shipping it. Each one comes with a way to tell whether it works: an eval that fails below its threshold, a test written after a real failure, a method written clearly enough that someone else could run it.
rolefit: reads job postings the way a careful reader would. Residency rules and years floors sit in the body, not the structured fields. Fixed checks first, a model's judgment on top, and an eval that fails below its threshold.
ai-landscape: a pipeline that turns what I read into published summaries, with the summarizing step graded. An invented number is caught by a fixed check; a dropped hedge or a flipped finding is caught by a model that's graded against test cases.
demo-runner: a product demo that runs and narrates itself, moving to the next scene when the narration ends instead of on a timer.