With the growing global need for housing and infrastructure, 3D concrete printing (3DCP) has emerged as an innovative construction method offering several potential benefits including design flexibility, speed, and sustainability. However, enhancing the reliability of 3DCP involves managing a variety of parameters that influence various aspects of the 3D printed structure. Process parameters like nozzle velocity, nozzle diameter, nozzle height, and material flow velocity have a major impact on the structural stability and filament shape. This project aimed to develop fast and accurate data-driven models for predicting and classifying filament shape based on process parameters. A print experiment systematically varied process parameters across 144 samples. The resulting filament geometry (width, height, contact width) was measured and classified by quality. Models were trained on this data to predict filament width, contact width, filament height, and classify filaments. These models can be utilized with any buildable material - a material with a high enough yield stress to bear the weight of upper layers without significant deformation. This condition does not restrict this study’s scope as it is a prerequisite for all 3DCP applications. The models’ robustness and generalizability were confirmed through validation on literature data across various printable materials and setups. These data-driven models can aid in optimizing parameters, generating variable width filaments, and printing non-planar layers. By linking print inputs to filament outputs, this comprehensive modeling approach advances 3DCP research for more reliable and versatile concrete printing.