Contact person: Business Client

Phone:Show

Email:Show

Location: India

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I’m looking to build a functional prototype of an AI-driven examination portal that focuses on lecture-based learning data. The goal is to prove the core concept end-to-end, so I can later scale it into a full SaaS product.

What I need you to develop:

• Lecture-to-Question Generator
Ingest a recorded lecture (audio or video), segment it by topic, and automatically create a mixed set of questions. Multiple-choice is welcome, but the must-have is auto-generated fill-in-the-blank items.

• Smart Evaluation
On submission, the system instantly checks those fill-in-the-blank answers with NLP similarity scoring or keyword matching, then stores the graded attempt.

• Adaptive Testing Logic
Serve the next question set based on the student’s current performance level so each learner gets a personalized path.

• Analytics Dashboard
For every completed test show:
– Accuracy analysis per topic
– Speed analysis per question and overall test
– Weak-area identification with visual charts (pie, bar, or line)
The admin view should also let me compare multiple tests on one screen.

• Lightweight AI Proctoring
Flag obvious cheating by monitoring webcam feed or screen focus changes during the attempt. A simple rules-based or model-based proof is enough for this phase.

User flow to cover:
1. Student watches a lecture; system logs watch duration, skips, and drop points.
2. The platform proposes a topic-wise test generated from that lecture.
3. Student takes the test and submits.
4. Immediate analytics appear; clicking any wrong question reveals a step-by-step solution and deep-links back to the exact lecture segment.
5. Admin reviews combined analytics across all learners.

Deliverables include:
• Source code (Python, Node, or comparable modern stack)
• Clear setup instructions and sample data
• Brief technical documentation explaining the AI models, evaluation logic, and future extension points

A modular, well-commented prototype that runs locally or on a small cloud instance will meet my needs for this milestone." (client-provided description)