This project investigates which biomechanical factors best predict counter movement jump height, specifically comparing impulse-based and power-based approaches, with the goal of giving athletes and coaches actionable training guidance. The work involved inheriting and auditing an existing multi-contributor code base, documenting all project folders, and building a physics-informed Lasso regression pipeline from the ground up using Python and scikit-learn. Through three generations of feature engineering and cross-validation refinement, two validated predictive models were produced and evaluated against a carefully constructed holdout set designed to test generalization across variable and unseen participants. Findings are being written up as a formal academic paper, which has involved drafting the introduction, methods, discussion, and conclusion sections, integrating figures directly from the analysis pipeline, coordinating with collaborators on the literature review and data collection sections, and iterating on the modeling narrative to accurately reflect each design decision made throughout the project.