active react typescript supabase ai client-work
Mi 70-20-10 — Leadership Coaching Platform
A coaching platform for a leadership-development program: submissions with AI-assisted evaluation, coach feedback, action plans, and an AI coach.
September 1, 2026
Problem
eQual Business Solutions’ leadership-development program was running on documents, spreadsheets, and email. Participants submitted work, coaches graded it against a rubric, and feedback lived in scattered files — slow for coaches and hard for anyone to follow.
Approach
I designed and built the platform end to end, and I run it in production.
- Front end — a React + TypeScript single-page app, built on a component library I extracted into its own design-system package.
- Backend — Postgres with row-level security, so participants, coaches, and admins each see exactly what their role allows. Role changes can’t be made from the client.
- AI — rubric-based evaluation of participant submissions, an AI coach chat, and retrieval over the program’s own material, with coaches deciding what feedback is released.
- Operations — admin tools to manage people and coach assignments, login history, a read-only “view as” mode for support, and calendar export for action items.
- Delivery — automated deploys for the app and backend functions on every push.
Outcome
The program now runs in one place: participants submit and track their plan, coaches review AI-assisted evaluations before releasing feedback, and admins configure the program without touching the database.