MyGooru
One reasoning engine. Five verticals. A stakeholder-facing layer I own end to end.
MyGooru is built on MAP (MyGooru AI Pathways) — reasoning infrastructure, developed over 13 years, that turns an organization's goals into personalized navigation for the people inside it. The same engine reasons about a student's next lesson, a shopfloor operator's next certification, or a government agency's next procurement step.
The product that reaches people is called the Companion — a role-adaptive interface that surfaces "the next right action" in context, either as a standalone app or embedded via API into an existing system. My job has been to design that stakeholder-facing layer across every vertical it ships in, education first and enterprise verticals since.
Every vertical needed a different job done. The engine underneath stayed the same.
When I picked up the design, the surfaces that used the reasoning engine had drifted from each other — different visual languages, different interaction patterns, no shared vocabulary for showing a user why the system recommended what it did. A teacher's dashboard and a plant manager's dashboard looked like they came from different companies, even though the same belief-and-pathway engine sat underneath both.
The harder problem was translation. The engine reasons in its own vocabulary — beliefs, confidence, pathways. A COO or a shopfloor supervisor doesn't think in that vocabulary. Every surface had to translate the same underlying reasoning into the language of the job the person in front of it was actually doing.
UI Design Lead across the shared system and every vertical surface.
I own the design system that all Companion surfaces pull from — tokens, layout patterns, the visual language for showing provenance (is this recommendation evidenced by real data, or inherited from a template, or a system hypothesis?) — and I design each vertical's surface on top of it: the K-12 education apps, and the enterprise-facing navigation used in manufacturing, government, and services deployments.
Where the work needs research — teacher workflow validation, plant supervisor interviews, government stakeholder reviews — I run it myself and fold findings back into the shared system, so a pattern learned in one vertical strengthens the others instead of forking them.
Shared design rules, vertical-specific surfaces.
Six stakeholder roles, one design language, real districts using it.
In education, MyGooru serves six roles: learners, instructors, institutional leaders, mentors (parents/guardians), curriculum developers, and IT administrators. The learner experience works like a GPS for mastery — it identifies gaps, delivers personalized resources, and tracks progress. Instructors get intervention timing and productivity tool recommendations; leaders get real-time data to coordinate strategy across schools.
The same engine, reasoning about a shopfloor, a procurement pipeline, a licensing queue.
The enterprise side of MyGooru takes the same MAP engine and points it at manufacturing competency, government workforce and procurement, and services readiness. The design problem was identical in shape to education: translate a reasoning engine into a vocabulary the person actually doing the job would recognize — an operator, a plant leader, an agency caseworker.
One design language, live across a platform reaching 5.3 million learners and multiple industrial and government deployments.
As many honest vocabularies as there are jobs to do.