Due to their speed and accuracy, computerized decision-making and data-analysis systems are often perceived to be close to the ideal of unbounded rationality. Challenging this perception, this study explores the possibility that computerized systems reflect the individual characteristics of the developers who have designed and realized them. Through the analysis of interviews with high-frequency trading system developers and a survey of software developers working in diverse industries, it shows that developers' characteristics, which have been related to bounded rationality (e.g. experience and expertise), influence their codes' performance and errors. Computer codes also reflect the personal circumstances of the developers (e.g. deadlines and stress), and their values. However, whereas some developers value code characteristics that are congruent with the ideal of unbounded rationality, including speed and accuracy, others aim to achieve characteristics that could be incongruent with it (e.g. code readability and modularity). Developers' roles and organizational practices, such as testing procedures and code reviews, limit the expression of their characteristics in their codes. Nevertheless, developers can be often identified by reading the codes. Highlighting that developers transfer some of their characteristics to their codes, this study identifies elements of bounded rationality in decisionmaking systems and extends organizational decision-making research.