" We solved the structural risk and viability problems of building a paid micro-learning product on third-party video by re-architecting it into a free-to-learn outbound-link skill tree backed by a paid Verified Credential, resulting in a fully specified, validation-gated platform with eliminated supplier dependency, defined compliance posture, and a defensible moat anchored in credential credibility rather than borrowed content. "
Overview
A free-to-learn skill-tree mobile and web platform with a paid Verified Credential as its revenue engine.
The product is designed as a dependency-locked, gamified learning path where each concept pairs an AI-generated study summary with a curated outbound link to a public resource and a server-graded quiz. Case study of the design and architectural work for a forthcoming development
Problem
Three failures in the existing learning landscape were targeted simultaneously: (1) public video platforms hold abundant educational content but offer no structure, sequencing, or active evaluation, leaving learners paralyzed by choice
(2) traditional EdTech bootcamps and courses carry costs that exclude most upskillers and suffer single-digit completion rates
(3) even learners who self-direct successfully have no verifiable proof to present to employers — free tools can teach but cannot certify. The combined effect is a market of motivated learners with no path to demonstrable, employer-recognized skill.
Challenges
-
Curriculum Sequencing
Transform open-ended job-role inputs into structured, dependency-aware learning pathways without losing logical progression. -
AI Content Reliability
Prevent AI-generated errors from affecting the quality and credibility of certification programs. -
Learning Path Validation
Ensure generated tracks maintain correct prerequisite relationships, skill progression, and curriculum integrity. -
Resource & Assessment Alignment
Match learning resources and quizzes accurately to generated skills while maintaining educational relevance. -
Quality Assurance
Introduce mandatory human review and approval before generated learning tracks are published to production.
Solution
The platform was delivered as an AI-assisted certification and learning-path ecosystem built around structured curriculum management, secure credential issuance, and dependency-driven learning progression. The architecture combines cross-platform delivery, server-side certification workflows, and a flexible content model designed for future specialization.
- Cross-Platform Learning Platform
The product architecture pairs a Flutter codebase (web + Android) with a PostgreSQL backend and Edge Functions for any server-only logic. - Structured Content Hierarchy
Content is modeled as a four-level hierarchy (Job Role → Section → Unit → Concept). - Dependency-Based Learning Progression
With dependency-locked unlocking at both Unit and Concept levels. - Future-Ready Content Architecture
A branch-ready schema that supports trunk-and-branch specialization without future migration. - Mobile OTP Authentication
Mobile OTP authentication is integrated with custom Supabase session minting to keep DLT-registered SMS templates in the developer's control. - Secure Credential Lifecycle
The credential follows a strict server-only state machine — LOCKED → ELIGIBLE → PAID → ASSESSMENT_STARTED → PASSED → ISSUED. - Certification & Retry Management
With re-pay on retry.
Architecture
The platform combines a Flutter frontend with a Supabase-powered backend to deliver a scalable learning and certification ecosystem. Structured curriculum generation, authentication, payments, assessments, and credential issuance are managed through server-side services and automated workflows.
Flutter (Dart) - Web & Android learning platform
Supabase - Authentication · Database · Storage · APIs
PostgreSQL - Job Roles · Sections · Units · Concepts · Certifications
Edge Functions - OTP · Grading · Payments · Credential Issuance
Razorpay & Play Billing - Subscription & certification payments
Curriculum Engine - DAG-based learning path generation & resource matching
Row-Level Security - Data protection & access control
Multi-Tenant Architecture - Scalable SaaS platform with isolated user access**
Key Features
- Outbound-link learning architecture that delivers structured study without embedding, cropping, or storing third-party video content — eliminating platform, IP, and creator-dependency risk.
- Dual-level dependency locking across Units and Concepts, enforcing deterministic progression while supporting future trunk-and-branch specialization through a branch-ready schema.
- AI-generated study summaries paired with curated outbound resources and timestamp cueing, delivered as the in-app learning layer.
- Server-graded quizzes with three formats (multiple choice, sequence reconstruction, diagnostic tap) and automated answer-key validation before publish.
- Gamified retention loop with streaks, XP, hearts, and a spaced-repetition schedule (SM-2 variant) for missed concepts.
- One-time Verified Credential with rigorous in-app proctored assessment, server-only state machine, public verification endpoint, and re-pay-on-retry policy designed to keep the credential scarce and meaningful.
- Dual-channel payments (Razorpay web + Play Billing / user-choice billing app) with cross-channel entitlement and no anti-steering violations.
- Admin panel with track creator, content moderation queues, business and content-health reports, learner data and progress views, and audit-logged impersonation for support.
- Compliance-by-design: full account deletion with credential voiding, mandatory up-front disclosure, DLT-registered SMS, mandatory Play Store and Razorpay pages, and an unavoidable minimal payment-record carve-out for legal retention.
- Phased build sequence designed so the free-learning loop (Phases 0–2) doubles as the validation vehicle before any payments engineering begins.