Artificial Intelligence, particularly in Machine Learning and related research areas such as Operational Research, currently faces a reproducibility crisis. Researchers encounter difficulties reproducing key results due to lacking critical details, including the disconnection between publications and the associated codes, data, and parameter settings. Solutions that improve code accessibility, data provenance tracking, research transparency, auditing of obtained results, and trust can significantly accelerate algorithm and model development, validation, and transition into real-world applications. Blockchain technology, with its features of decentralization, data immutability, cryptographic hash functions, and consensus algorithms, provides a promising avenue for developing such solutions. By leveraging the distributed ledger working over a peer-to-peer network, a secure and auditable infrastructure can be established for sharing and controlling data in a trusted manner. Our analysis examines the current state-of-the-art blockchain-based proposals that target reproducibility issues in the Machine Learning domain. Based on the analysis of existing solutions, we propose a high-level architecture and main modules for developing a blockchain-based platform that enhances reproducible research in Machine Learning and can be adapted to other Artificial Intelligence domains.