You are an expert full-stack developer and software architect. Build the first MVP version of a Personalized Education Web Application using Laravel 13 and MySQL.

The application architecture must be API-first so that it can easily support mobile applications (iOS/Android) in the future, while providing a clean web dashboard for data management.

1. CORE CONCEPT & ROLES

The core of the system is the "Student Profile". The platform serves three main user roles:

• Teacher: Can view/edit student profiles, add progress logs, and trigger AI lesson generation.
• Parent: Can view their child's profile and update home-based progress logs.
• Student: Can view their own profile and access generated exercises.
• Admin: Full system management.

2. DATABASE SCHEMA (MySQL)

Generate Laravel migrations with proper foreign keys, indexes, and soft deletes for the following tables:

• users: id, name, email, password, role (enum: admin, teacher, parent, student), timestamps, softDeletes.
• curriculum_blocks: id, name (string), description (text), start_date (date), end_date (date), timestamps, softDeletes.
• student_profiles: id, user_id (foreign key to users, unique), vision (text), mission (text), education_method (enum: waldorf, montessori, stem, traditional), interests (text), ai_generation_status (enum: idle, generating, failed), timestamps, softDeletes.
• student_relations: id, student_profile_id (foreign key), user_id (foreign key to users), relation_type (enum: parent, teacher), timestamps.
• capabilities: id, student_profile_id (foreign key), skill_name (string), current_level (text), notes (text), timestamps.
• progress_logs: id, student_profile_id (foreign key), updated_by (foreign key to users), content (text), logs_date (date), timestamps, softDeletes.
• personalization_state:
  • id
  • student_profile_id (foreign key to student_profiles, unique)
  • ai_condensed_memory (json or text - AI-synthesized cumulative analysis of the student's psychological & developmental state)
  • learning_style_nuances (text - behavioral traits, e.g., "responds well to physical movement/rhythm")
  • current_bottlenecks (text - core challenges the student is facing)
  • last_processed_log_id (foreign key to progress_logs, nullable - tracks up to which log the AI has already summarized)
  • timestamps
• suggested_exercises: id, student_profile_id (foreign key), curriculum_block_id (foreign key, nullable), title (string), description (text), expected_outcome (text), status (enum: pending, active, archived), timestamps, softDeletes.
• exercise_submissions: id, suggested_exercise_id (foreign key), student_notes (text), teacher_evaluation (text), status (enum: submitted, reviewed), timestamps.
• media_attachments: id, attachable_id (integer), attachable_type (string for polymorphic relation to support progress_logs and exercise_submissions), file_path (string), file_type (enum: image, video, document), timestamps.

3. BACKEND ARCHITECTURE (Laravel 13)

Please implement the following components:

• Authentication: Setup Sanctum or standard token-based authentication for secure API endpoints.
• Controllers & Routes:
  • StudentProfileController: CRUD operations for profiles, filtered by user permissions.
  • ProgressLogController: Allow teachers/parents to post daily/weekly learning logs.
  • AIExerciseController: Trigger the AI engine to generate localized, method-specific exercises.
• AI Integration Service (AIEngineService):
  • Create a service that connects to an external Ollama/OpenAI-compatible endpoint via Illuminate\Support\Facades\Http.
  • Fetch the endpoint from .env (OLLAMA_API_URL and OLLAMA_MODEL).
  • Read student profile data (method, capabilities, interests, vision) and send a structured prompt to the AI.
  • The AI must output pure JSON matching the suggested_exercises table schema.
• Background Jobs: Use Laravel Queue (ShouldQueue) for the AI integration to process LLM API requests asynchronously and avoid timing out web requests.

4. EDUCATION METHOD LOGIC (Crucial for AI Prompting)

Inside the AIEngineService, the prompt construction must respect the chosen education_method:

• If waldorf: Focus tasks on experiential learning, natural materials, storytelling, and arts.
• If montessori: Focus tasks on self-directed activity, practical life skills, and hands-on learning.
• If stem: Focus tasks on logic, problem-solving, and technical integration.

5. FRONTEND INTERFACE (Simple & Modern)

• Provide a clean, modern web interface using Tailwind CSS (via Blade or Inertia.js) to demonstrate the Student Profile dashboard, where a teacher can click a "Generate AI Learning Path" button and see the results update dynamically via AJAX/API.

OUTPUT EXPECTATION

Please generate all necessary Migrations, Models (with relationships defined), Controllers, Services, Jobs, API/Web Routes, and a basic Dashboard View for the Student Profile. Ensure clean code, proper validation, and error handling (Try-Catch blocks for the AI API calls).