AI-Based C++ Code Generation for Fitting Models in FITTER_WEB

Abstract The article discusses the possibilities of using artificial intelligence (AI) to automate the process of generating code in C++ in the FITTER_WEB web application designed for fitting experimental small-angle scattering data. The architecture of integrating the DeepSeek language model into the system is presented, which allows to significantly increase the speed and quality of preparing models for data analysis. A comparison of the code written by the programmer and the AI for some models is carried out.