Learning Management System

This isn't a plain course-and-video LMS. Content is organized as "Playbooks" — structured programs with sections and lessons that mix video, audio and article formats — sitting alongside a goal-setting module (users define a personal "why," set goals, log actions and track progress), scheduled live team classes with assigned mentors, a quiz/assessment engine with per-category results, and a gamification layer of badges and login/exercise streaks.

The most technically interesting piece is the AI layer. An in-app assistant lets a user chat about their content and ask for a summary of the specific lesson they're on; the request carries the exact playbook/section/lesson IDs so the backend can ground its answer in that content. That backend isn't a single API call — it's a TypeScript Firebase Cloud Functions layer (chat, recommendations, an audio processor, a daily-exercise picker) that itself calls out to OpenAI, Google's Gemini SDK and Groq, plus a separate Python/Flask service that builds a Pinecone vector index over the course library for retrieval and runs an agent-based assessment pipeline.

The student app and admin console share one Firestore backend and Riverpod for state on both sides, so a course or announcement an admin publishes is live for students without a redeploy. The admin console itself is role-aware: full admins get playbook/user/mentor/team management, dashboards and content editors, while coaches get a scoped view — their teams, their team's members, and notifications — behind the same login screen.

Monetization runs through RevenueCat (in-app purchases/subscriptions with a trial flow), and premium lesson content is protected against screen capture. The student app also ships offline-first touches (Hive-cached data, a background audio player with lock-screen controls for audio lessons) and push notifications via Firebase Cloud Messaging.