AI Exam Revision Assistant

I’m building a professional AI-driven revision application with regular subscription aimed specifically at healthcare and medicine professional certification exams. The core idea is simple: I supply the source material—textbooks, slide decks, clinical guidelines, and banks of past exam questions—then the system learns from that content, mirrors the style of those questions, and continuously generates fresh practice items.

Key workflow • As an admin I can upload or update source documents and historical questions; the system re-indexes them automatically. • Learners interact through a clean text-based chat. The AI answers content questions, serves up practice quizzes, and explains solutions. • Behind the scenes, the model tracks performance, pinpoints weak topics, and adapts each subsequent quiz to target those gaps until mastery is demonstrated. • All questions it generates must match the structure, cognitive level, and wording style of the originals so users feel they are sitting a realistic mock exam.

Deliverables

1. Data-ingestion module with simple UI for uploading and tagging PDFs, Word docs, or CSV question sets.
2. Fine-tuned question-generation model (OpenAI, GPT-J, or similar) that reproduces the original format and difficulty spectrum.
3. Adaptive engine that analyses per-topic accuracy and dynamically selects or creates new questions.
4. Chat interface (React, Flutter, or similar front end; Node/Python back end) that supports explanation threads, progress analytics, and bookmarking of tricky questions.
5. Admin dashboard for cohort statistics and content management.
6. Deployment scripts and brief technical documentation so I can run it on AWS or Azure.

Acceptance criteria • Generated questions must achieve ≥90 % stylistic similarity in a blind review against real past papers. • Adaptive algorithm demonstrably increases a pilot user’s weak-topic score by at least 20 % after three sessions. • Source uploads trigger re-training or re-indexing within 10 minutes without downtime.

If you’ve previously built educational chatbots, fine-tuned large language models, or implemented spaced-repetition/adaptive learning logic, I’d love to see your approach and a short timeline. Budget: EUR 250–750 Skills: Python, Machine Learning (ML), Data Science, Data Analysis, Data Management, Natural Language Processing, AI Chatbot, AI Model Development