A Smart Peer-Learning Platform with AI-Driven Skill Matching and a Credit-Based Reciprocal Exchange Mechanism

Abstract: Peer-to-peer learning lets individuals teach skills they possess and learn skills they lack, but informal arrangements suffer from poor discoverability, unbalanced reciprocity, and a lack of trust between strangers. Existing online learning platforms are predominantly one-directional and monetary, offering little support for equitable, non-cash exchange among peers. This paper presents a smart peer-learning platform that pairs an artificial-intelligence skill- matching recommender with a credit-based trading mechanism, enabling learners to earn credits by teaching and spend them to receive instruction. A hybrid recommender combines content-based skill similarity with collaborative signals to suggest high-quality learner-mentor pairs, a reputation engine aggregates peer ratings to build trust, and a credit ledger with escrow guarantees fair settlement of each session. A Python back end implements the recommendation, reputation, and ledger logic, while a Node.js layer delivers the marketplace and session interfaces. Evaluated against popularity, content-based, and collaborative-filtering baselines, the hybrid recommender achieved a precision-at-five of 0.88 and a recall-at-five of 0.85, and simulated platform activity exhibited sustained growth in completed sessions and circulated credits. The principal contributions are a hybrid skill-matching recommender tailored to bidirectional learning, a credit- with-escrow exchange protocol that enforces reciprocity, and a reputation mechanism that fosters trust in an open peer marketplace.

Keywords: Peer learning; recommender systems; skill matching; credit-based exchange; reputation systems; collaborative learning; hybrid recommendation; online education.