Analysis of intelligent jumbo measurement-while-drilling(MWD) parameters reveals that drilling speed stands as the most scientifically valid and operationally accessible indicator for assessing in-situ rock mass conditions prior to tunnel excavation. Analysis of 3732 peripheral holes across 84 blasting rounds from an intelligent jumbo in high-altitude tunneling revealed systematic measurement-while-drilling(MWD) datasets that quantify correlations between penetration rate, explosive parameters, and excavation overbreak/underbreak. The developed analytical framework incorporates four key procedures: identification of pilot drilling phases; clustering analysis of steady-state penetration rates; vectorization and dimensional reduction of charge configurations; all derived by incorporating geometric constraints into the computation of the minimum burden. The extracted parameters comprise peripheral hole extrapolation angles, hole spacing, minimum resistance line, explosive charge configurations, and drilling speed. The MWD characteristics serve as inputs for developing a LightGBM-based excavation-contour prediction system. This computational framework generates quantitative estimates of design-induced overbreak and underbreak dimensions, incorporating early-termination protocols and Bayesian-optimized hyperparameters to enhance model convergence and predictive generalization. Experimental results demonstrate effective model convergence during training, with validation/testing errors maintained within practical engineering tolerances. The model successfully characterizes overbreak-underbreak variations that are correlated with minimum burden, charging configurations, and drilling-speed features. Field implementation achieved significant improvements: average linear overbreak decreased by 46.8% (from 24.8 cm to 13.2 cm), and the half-cast factor increased by 49.1% (from 53% to 79%), demonstrating enhanced contour precision and blast fragmentation quality.