AI-Driven IELTS Prep Platform

Build a modern, high-performing web application that lets IELTS candidates practise every skill while receiving instant, AI-generated feedback and score predictions for Speaking and Writing. The core platform must:

• Accept Speaking and Writing submissions, run them through an AI model, and return band-style scores with actionable feedback.
• Let administrators upload and manage complete Reading and Listening practice tests (audio, passages, answer keys, timers).

Analytics & dashboards
Students will log in to a personalised dashboard that visualises their progress over time, while a separate administrator view aggregates cohort performance and drills down to detailed test-by-test analytics. Key metrics include overall band trajectory, section strengths/weaknesses, question-level accuracy, and usage statistics.

Local payment integration
Monetisation runs through Uzbekistan’s Click and Payme gateways. The build must support subscription-style recurring charges as well as automated refund processing, all reflected in a transaction log inside the admin panel.

Technical expectations & deliverables
1. Responsive web app (desktop & mobile) built with a contemporary stack you propose.
2. Secure user authentication, role-based access (student, admin).
3. AI evaluation microservice or API wrapper that can be swapped or fine-tuned later.
4. Content management area for test uploads, publication scheduling, and score key entry.
5. Dashboards for student progress tracking, administrator performance overview, and detailed test-result analysis.
6. Full payment workflow with Click/Payme sandbox and production modes, covering recurring billing and refunds.
7. Deployment scripts and brief hand-over docs so I can spin up new environments independently.

Acceptance criteria
• Speaking/Writing submissions return a score within 10 seconds 95 % of the time.
• Payment callbacks are logged and reconciled with user entitlements in real time.
• All charts load in