Version 1.5 - February 2026

Mock Interview System

• Candidate Mock Interviews: Practice interviews with AI interviewer for real job postings before the actual interview
• Resume Session Support: Pause and resume incomplete mock interviews with time tracking
• AI-Powered Feedback: Receive detailed performance analysis, strengths, areas for improvement, and recommendations
• Credit System: 1 credit per mock interview session with configurable pricing

Mobile Responsiveness

• Mock Interview Room: Full mobile responsive design with touch-friendly controls (44px+ touch targets)
• Portrait/Landscape Support: Optimized layouts for both mobile orientations
• iOS Smooth Scrolling: Added -webkit-overflow-scrolling: touch for native scroll feel
• Extra Small Devices: Additional breakpoints at 400px for smaller screens
• Take Exam Page: Scrollable and responsive exam interface
• Interview Report: Mobile-optimized feedback and transcript display

Bulawork AI Assistant Improvements

• Real-time Database Integration: AI assistant now uses live data from 19 data sections including:
  • Active Jobs & Application Metrics
  • AI Interview Data & Pipeline Analytics
  • Skills Analytics & Recent Hires
  • Job Performance Metrics & Bottleneck Analysis
• Markdown Rendering: Clean response formatting with proper HTML rendering (no raw ##, **, - symbols)
• Data Integrity: Strict rules preventing fake/placeholder data - only real database values
• Branding: Renamed from "AI HR Assistant" to "Bulawork AI"

Performance & Database

• API Timeout Fix: Increased cURL timeout from 60 to 180 seconds for long interview analysis
• Token Limits: Added max_tokens: 4000 for AI responses to prevent overflows
• Transcript Truncation: Long interview transcripts (15,000+ chars) now truncated for processing
• Deterministic AI Analysis: Set temperature: 0 for consistent re-analysis results
• 30-Day Chat Cleanup: Automatic deletion of old AI chat history via cron job

Exam Creator Improvements

• Job Skills Integration: Skills now fetched from job_skills table with proper fallbacks
• Enhanced Data Retrieval: Falls back to required_skills, requirements, and description fields
• Responsibilities Extraction: Automatically extracts responsibilities from job descriptions when not explicitly set

AI Interview Analysis

• Removed Candidate Ranking: Eliminated unrealistic percentile rankings when no comparison data exists
• Incomplete Interview Detection: Low scores (15%) for interviews under 1 minute with minimal responses
• Fair Scoring: Context-sensitive ratings based on actual interview content

Files Modified

Key Features

Technology Stack

Core Database Tables

User Management (7 tables)

Company Management (8 tables)

Jobs & Applications (10 tables)

AI Features (8 tables)

Exam System (7 tables)

Security Tables (6 tables)

Authentication APIs

Location: /api/auth/

// Request
{
    "email": "user@example.com",
    "password": "securepassword",
    "remember": true,
    "trust_device": false
}

// Response (Success)
{
    "success": true,
    "requires_otp": false,
    "redirect_url": "/hr/dashboard.php",
    "message": "Welcome back, John!"
}

// Response (OTP Required)
{
    "success": true,
    "requires_otp": true,
    "email": "user@example.com"
}

Candidate APIs

Location: /api/candidate/ - Requires authenticated candidate session

HR APIs

Location: /api/hr/ - Requires authenticated HR session

Admin APIs

Location: /api/admin/ - Requires admin authentication

Mobile APIs

Location: /api/mobile/ - Token-based authentication for native apps

Implemented Security Features

Authentication Flow

Document Encryption System

All sensitive documents (resumes, qualifications, application attachments) are encrypted using AES-256-GCM before storage.

// Encryption Process (config/encryption.php)
1. Generate unique document ID for key derivation
2. Derive document-specific key using HKDF from master key
3. Generate random 12-byte IV (nonce)
4. Encrypt with AES-256-GCM (authenticated encryption)
5. Store: [IV (12 bytes)] + [Tag (16 bytes)] + [Ciphertext]
6. Add .enc extension to filename

// Decryption Process
1. Extract IV, Tag, and Ciphertext from file
2. Derive document key using stored document ID
3. Decrypt and verify authentication tag
4. Return plaintext only if authentication passes

CSS Design System

BulaWork uses a custom CSS design system with CSS variables for consistent theming.

Core Color Palette

:root {
    /* Primary Colors */
    --primary-dark: #2F2E2E;
    --charcoal: #36454F;
    --charcoal-dark: #2D3748;
    --accent-yellow: #F3D15F;
    --accent-yellow-hover: #E5C34D;
    
    /* Background Colors */
    --bg-cream: #F9FAFB;
    --bg-dashboard: linear-gradient(135deg, #F9FAFB 0%, #FFFBEB 40%, #FDE047 100%);
    --card-white: #FFFFFF;
    
    /* Status Colors */
    --success: #10B981;
    --info: #3B82F6;
    --warning: #F59E0B;
    --danger: #EF4444;
    --purple: #8B5CF6;
    
    /* Border Radius */
    --border-radius-xl: 32px;
    --border-radius-lg: 24px;
    --border-radius-md: 16px;
    --border-radius-sm: 12px;
    
    /* Shadows */
    --shadow-soft: 0 10px 40px -10px rgba(47, 46, 46, 0.10);
    --shadow-hover: 0 20px 50px -10px rgba(156, 141, 112, 0.25);
    --shadow-charcoal: 0 8px 25px rgba(54, 69, 79, 0.25);
}

CSS Files Structure

JavaScript Modules

Frontend JavaScript is organized into modular files for different functionalities.

Key JavaScript Patterns

// AJAX Form Submission Pattern
async function submitForm(formData, endpoint) {
    try {
        const response = await fetch(endpoint, {
            method: 'POST',
            headers: { 'X-Requested-With': 'XMLHttpRequest' },
            body: formData
        });
        const data = await response.json();
        if (data.success) {
            showToast('Success!', 'success');
        } else {
            showToast(data.error || 'Error occurred', 'error');
        }
    } catch (error) {
        showToast('Network error', 'error');
    }
}

// Modal Pattern
function openModal(modalId) {
    document.getElementById(modalId).classList.add('active');
    document.body.style.overflow = 'hidden';
}

function closeModal(modalId) {
    document.getElementById(modalId).classList.remove('active');
    document.body.style.overflow = '';
}

Server Requirements

• PHP 8.0 or higher with extensions: PDO, OpenSSL, JSON, mbstring, fileinfo
• MySQL 8.0 or MariaDB 10.5+
• Apache or Nginx web server with mod_rewrite
• HTTPS certificate (required for production)
• Minimum 512MB RAM, 5GB storage

Installation Steps

1. Clone repository to web server document root
2. Create MySQL database and import database/bulawork_deployment_v4.sql
3. Configure config/database.php with database credentials
4. Set file permissions: chmod 755 for directories, 644 for files
5. Create writable uploads directory: chmod 775 uploads/
6. Configure SMTP settings via Admin > Settings > Email
7. Set OpenAI API key in config/api.php for AI features
8. Create first admin user manually in database or via CLI

Cron Jobs

# Job alert emails - every 6 hours
0 */6 * * * php /path/to/bulaw/cron/job-alerts.php

# Friday HR summary - Fridays at 5 PM
0 17 * * 5 php /path/to/bulaw/cron/friday-summary.php

# Archive cleanup - daily at 2 AM
0 2 * * * php /path/to/bulaw/cron/archive-cleanup.php

# Saved jobs expiry check - daily at 6 AM
0 6 * * * php /path/to/bulaw/cron/saved-jobs-expiry.php

# AI Chat history cleanup (30-day retention) - daily at 3 AM
0 3 * * * php /path/to/bulaw/cron/chat-cleanup.php