Aim: This study aims to assess the applicability of sparse Gaussian process regression (sGPR) modeling to analyze 12-lead ECG waveforms to assess likelihood of Chagas disease, as an entry to the George B. Moody Physionet Challenge 2025. Methods: We propose to do this via multi-output, sparse GPR methods. Gaussian process regression is a valuable tool for uncertainty quantification in multivariate regression. Sparse GPR, with fewer basis functions, is less computationally expensive than standard GPR, and enables faster training over the larger datasets. The preliminary training of the sGPR models, with a five-fold cross-validation, is being done on the CODE-15% dataset to extract features from the 12-lead ECG waveforms that may potentially be indicative of Chagas disease. The features are fed into a logistic regression model to rank their impact on the predictions on a subset of SaMi-Trop dataset (the "True-Positive" dataset). The prominent features are then validated against the PTB-XL and a set-aside subset of SaMi-Trop datasets to determine the features that are prominent in the true-positive vs true-negative dataset. These features are then used to evaluate self-reported Chagas disease in the CODE-15% dataset. Results: Our initial model, which was trained on a limited part of CODE-15% dataset (9,990 waveforms), had an Area Under the Receiver Operating Characteristic of 0.5, and an Area Under the Precision-Recall Curve of 0.011, with an accuracy of 0.989, during internal validations. Due to this limited training, our preliminary model, while being successfully evaluated on the validation dataset in the unofficial phase of the Challenge, performed poorly, scoring 0.037 on the validation data. Conclusion: Sparse Gaussian process regression can yield interesting perspectives on the potential of using 12-lead ECG waveforms from individual patients to predict Chagas disease. Identifying uncertainty in the predicted likelihood can help optimize the outlay of serological testing for such patients.