<h2>The Problem</h2>
<p>Aspiring architects in the Philippines face a difficult board exam: the Architect Licensure Examination (ALE), administered by the Professional Regulation Commission (PRC). Pass rates hover around 40-60%. Yet study resources are scattered. Reviewers rely on outdated photocopied materials, expensive review centers, or random question compilations shared in Facebook groups.</p>
<p>There was no centralized, modern platform where examinees could practice with structured questions, track their weak areas, and study efficiently. The gap between available tools and what students actually needed was enormous.</p>
<h2>What I Built</h2>
<p><a href="https://www.arkitekto.review">Arkitekto</a> is a comprehensive online review platform covering all ALE exam subjects. It provides 7,000+ practice questions, mock examinations, spaced repetition review, progress tracking, study groups, and community discussions.</p>
<h3>Core Features</h3>
<ul>
<li><strong>Practice Mode</strong>: Study by subject with immediate feedback and explanations</li>
<li><strong>Mock Examinations</strong>: Timed exams simulating the actual ALE format with scoring and review</li>
<li><strong>Spaced Repetition Review</strong>: SM-2 algorithm schedules questions based on retention, similar to Anki</li>
<li><strong>Weak Area Practice</strong>: Focus sessions targeting subjects where the user struggles most</li>
<li><strong>Flashcard Mode</strong>: Quick review of questions in a flashcard format</li>
<li><strong>Progress Tracking</strong>: Study streaks, subject mastery levels, performance analytics, and achievement badges</li>
<li><strong>Study Groups</strong>: Collaborative study spaces where users can practice together</li>
<li><strong>Discussions</strong>: Community Q&#x26;A threads linked to specific questions</li>
<li><strong>Bookmarks</strong>: Save difficult questions for later review</li>
<li><strong>Admin Dashboard</strong>: Content moderation, user management, question management, and analytics</li>
</ul>
<h2>Tech Stack</h2>
<table>
<thead>
<tr>
<th>Layer</th>
<th>Technology</th>
<th>Why I Chose It</th>
</tr>
</thead>
<tbody>
<tr>
<td>Framework</td>
<td><a href="https://nextjs.org/">Next.js 16</a></td>
<td>App Router, Server Components, React 19</td>
</tr>
<tr>
<td>Database</td>
<td><a href="https://supabase.com/">Supabase</a></td>
<td>PostgreSQL with RLS, Auth, real-time</td>
</tr>
<tr>
<td>Styling</td>
<td><a href="https://tailwindcss.com/">Tailwind CSS 4</a></td>
<td>Utility first, design tokens</td>
</tr>
<tr>
<td>Components</td>
<td><a href="https://ui.shadcn.com/">shadcn/ui</a></td>
<td>Accessible, customizable primitives</td>
</tr>
<tr>
<td>State</td>
<td><a href="https://zustand-demo.pmnd.rs/">Zustand</a></td>
<td>Lightweight client state</td>
</tr>
<tr>
<td>Caching</td>
<td><a href="https://tanstack.com/query">React Query</a></td>
<td>Server state and cache management</td>
</tr>
<tr>
<td>Validation</td>
<td><a href="https://zod.dev/">Zod</a></td>
<td>Runtime schema validation</td>
</tr>
<tr>
<td>Editor</td>
<td><a href="https://lexical.dev/">Lexical</a></td>
<td>Rich text for bookmark notes</td>
</tr>
<tr>
<td>Hosting</td>
<td><a href="https://vercel.com/">Vercel</a></td>
<td>Edge caching, fast deploys</td>
</tr>
</tbody>
</table>
<h2>Architecture</h2>
<pre><code>flowchart TB
    subgraph Client["Frontend (Next.js 16 + React 19)"]
        SC[Server Components]
        CC[Client Components]
        SA[Server Actions]
        ZS[Zustand Store]
    end

    subgraph Auth["Authentication"]
        SBA[Supabase Auth]
        RLS[Row Level Security]
    end

    subgraph Database["Supabase PostgreSQL"]
        QV[sc_questions_full View]
        UP[sc_user_profiles]
        UQ[sc_user_question_stats]
        SR[sc_spaced_review]
        SG[sc_study_groups]
        EX[sc_exams]
    end

    subgraph Algorithms["Business Logic"]
        SM2[SM-2 Spaced Repetition]
        RS[Readiness Score]
        PERF[Performance Analytics]
    end

    SC --> SBA
    SA --> RLS
    RLS --> Database
    CC --> ZS
    SA --> Algorithms
    Algorithms --> Database
</code></pre>
<h3>Data Flow: Spaced Repetition</h3>
<pre><code>sequenceDiagram
    participant U as User
    participant F as Frontend
    participant SA as Server Action
    participant SM as SM-2 Algorithm
    participant DB as Supabase

    U->>F: Answer Question
    F->>SA: Submit Answer
    SA->>DB: Fetch Current Stats
    DB-->>SA: ease_factor, interval, repetitions
    SA->>SM: Calculate Next Review
    SM-->>SA: new_interval, new_ease_factor
    SA->>DB: Update sc_spaced_review
    SA->>DB: Update sc_user_question_stats
    SA-->>F: Result + Next Review Date
    F-->>U: Show Feedback + Explanation
</code></pre>
<h3>Question Schema</h3>
<pre><code>erDiagram
    sc_subjects ||--o{ sc_topics : contains
    sc_topics ||--o{ sc_questions : has
    sc_questions ||--o{ sc_question_options : has
    sc_questions ||--o{ sc_user_question_stats : tracked_by
    sc_questions ||--o{ sc_spaced_review : scheduled_in
    sc_questions ||--o{ sc_bookmarks : saved_in
    sc_questions ||--o{ sc_discussions : discussed_in

    sc_subjects {
        uuid id PK
        string name
        string code
    }

    sc_topics {
        uuid id PK
        uuid subject_id FK
        string name
    }

    sc_questions {
        uuid id PK
        uuid topic_id FK
        text question_text
        text explanation
        enum difficulty
    }

    sc_user_question_stats {
        uuid id PK
        uuid user_id FK
        uuid question_id FK
        int times_answered
        int times_correct
    }

    sc_spaced_review {
        uuid id PK
        uuid user_id FK
        uuid question_id FK
        float ease_factor
        int interval
        date next_review
    }
</code></pre>
<h2>What Building This Taught Me</h2>
<h3>1. Content quality is everything in EdTech</h3>
<p>I started with 7,000+ questions and thought the platform was ready. Then users started reporting inaccuracies. A wrong answer key on one question destroys trust in the entire platform.</p>
<p>I ended up running 12 rounds of automated and manual review across the entire question bank. AI-assisted review helped catch issues at scale, but every flagged question still needed human verification against authoritative sources like the National Building Code and RA 9266.</p>
<p>The lesson: in education, a smaller set of verified questions is more valuable than a large set of unverified ones.</p>
<h3>2. Spaced repetition needs careful tuning</h3>
<p>The SM-2 algorithm is well documented, but implementing it for exam preparation is different from language learning. Architecture exam questions aren't binary recall. They require understanding concepts across multiple subjects.</p>
<p>I had to adjust intervals and ease factors to account for the fact that missing a question about structural design doesn't mean you forgot it. You might have misunderstood the question or confused similar building codes.</p>
<p>Resources:</p>
<ul>
<li><a href="https://www.supermemo.com/en/archives1990-2015/english/ol/sm2">SM-2 Algorithm</a></li>
<li><a href="https://docs.ankiweb.net/studying.html">Anki's Implementation</a></li>
</ul>
<h3>3. Row Level Security changes how you think about data</h3>
<p>Supabase RLS means every query is automatically scoped to the authenticated user. This is powerful, but it forces you to design your schema around access patterns from day one.</p>
<p>Every table needs a <code>user_id</code> column. Every policy needs to be tested. And when you need cross-user queries (like study group leaderboards), you need database functions that bypass RLS with <code>SECURITY DEFINER</code>.</p>
<pre><code>-- RLS policy: users can only see their own stats
CREATE POLICY "Users can view own stats"
  ON sc_user_question_stats
  FOR SELECT
  USING (user_id = auth.uid());
</code></pre>
<h3>4. Niche audiences are deeply engaged</h3>
<p>Building for Filipino architecture students means a small but passionate user base. Every feature request comes with genuine context. Users tell you exactly what they need because the exam is real and the stakes are high.</p>
<p>This taught me that building for a niche is easier than building for everyone. The feedback loop is tighter, the community is supportive, and feature prioritization is obvious because the goal is clear: pass the board exam.</p>
<h2>The Bigger Realization</h2>
<p>Building Arkitekto taught me that EdTech isn't about technology. The hardest problems were content quality, pedagogical effectiveness, and understanding how people actually study.</p>
<p>I could have built a technically simpler platform with basic quizzes and it would have served the same purpose. The spaced repetition, analytics, and study groups matter, but only because the underlying content is accurate and comprehensive. Technology amplifies good content. It can't replace it.</p>
<h2>What I'd Do Differently</h2>
<p>Invest in content review infrastructure from the very beginning. I built the platform first and treated content quality as a later concern. That meant months of retroactive review scripts and manual verification.</p>
<p>If I started over, I'd build a review pipeline first: a system where every question goes through automated checks and human verification before it ever reaches a student.</p>
<h2>References</h2>
<ul>
<li><a href="https://nextjs.org/docs/app">Next.js App Router</a></li>
<li><a href="https://supabase.com/docs">Supabase Documentation</a></li>
<li><a href="https://www.supermemo.com/en/archives1990-2015/english/ol/sm2">SM-2 Spaced Repetition Algorithm</a></li>
<li><a href="https://www.officialgazette.gov.ph/2004/03/17/republic-act-no-9266/">Philippine RA 9266 - Architecture Act of 2004</a></li>
<li><a href="https://www.prc.gov.ph/">PRBOA Syllabus</a></li>
</ul>
<h2>Links</h2>
<ul>
<li><strong>Website:</strong> <a href="https://www.arkitekto.review">arkitekto.review</a></li>
</ul>


Projects.

Arkitekto: Philippine Architect Board Exam Reviewer