This study proposes an automated computer vision (CV) system to assess culinary task-related awkward postures. Traditional methods like observation and self-reports are subjective and time-consuming. To address this, the study applies CV and statistical tools to streamline ergonomic assessments in kitchens. It involved 34 participants–both laypersons and professionals–performing repetitive culinary tasks. Posture estimation was used to track joint positions and angles, evaluated using the Assessment of Repetitive Tasks (ART) method. Twelve tasks were analyzed: chopping, grating, lifting a pot, mixing, peeling, reaching into an oven, rolling, sawing, sautéing, slicing, stirring, and whisking. The automated system’s results were compared with manual evaluations from one certified ergonomist and three trained raters. Professionals showed greater agreement (r = 0.87) than laypersons (r = 0.81) according to a Fisher’s Z-test, Z = 4.88, p < 0.001. The findings confirm that the system assesses posture dependably, especially for professional cooks. This tool could improve the evaluation of ergonomics and support safer kitchen practices for both professionals and home chefs.