Abstract
Evaluation of scientific and scientific-technical projects is a multifaceted task that requires a balanced approach to the analysis of innovativeness, feasibility and potential impact of research initiatives. Traditional methods of expert evaluation are based on subjective conclusions of specialists, which create risks of inconsistency and limited reproducibility of solutions. In this regard, the importance of intelligent information systems that can improve the quality, transparency and efficiency of evaluation procedures is increasing. The article presents a comprehensive review of architectural solutions, network models and technologies used in information systems designed to support expert evaluation of scientific projects. Both classical knowledge-based approaches and modern solutions using artificial intelligence methods are considered: machine learning, natural language processing (NLP), fuzzy logic, and genetic algorithms. Particular attention is paid to the use of blockchain technologies to ensure the reliability, security and auditability of expert procedures. Key decision support methods are analyzed, including multi-criteria analysis methods, optimization algorithms, and automated distribution of applications between experts. Examples of successful implementation of cloud and distributed architectures in the scientific expertise system are presented.