Peer-to-peer (P2P) learning allows people to share skills directly with other learners instead of depending only on formal courses. Existing skill exchange platforms make this idea practical, but many of them still use keyword search or simple category matching. Such matching can miss useful relationships between related skills, users, learning goals, and previous interactions. This paper presents a proposed graph-based P2P skill exchange platform that uses Neo4j to model users, skills, interests, communities, sessions, and ratings as connected data. A graph-based recommendation process is then used to identify learning partners using skill similarity, shared interests, learning history, trust, ratings, and community participation. The work is based on a review of recent studies on skill exchange, knowledge graphs, graph learning, graph neural networks, and Neo4j-based recommendation. The main contribution is a unified framework that brings reciprocal learning and relationship-aware recommendation into one system. The paper focuses on system design and analytical discussion; it does not claim experimental accuracy because the proposed platform has not yet been experimentally implemented.

Parth G. Gaikwad, Srushti A. Rajkule, Shreya Kolam, Ganesh S. Patinge, Rucha Thakre, "Graph-Based Peer-to-Peer Skill Exchange Platform" International Journal of Scientific Research and Engineering Development, V9(5): Page(2174-2181) September - October 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.