Data-driven Chagas disease screening via electrocardiograms (ECGs) is limited by scarce and noisy labels in existing datasets. We propose a biomarker-based pretraining approach, where an ECG feature extractor is first trained to predict percentile-binned blood biomarkers from the MIMIC-IV-ECG dataset, using a bin-smoothing regularization to handle the sparsity introduced by binning. The pretrained model is then fine-tuned on Brazilian datasets (CODE15% and SaMi-Trop) for Chagas detection. Our 5-model ensemble, developed by the Ahus AIM team, achieved a challenge score of 0.412 on the hidden validation set, ranking 5/66 in Detection of Chagas Disease from the ECG: The George B. Moody PhysioNet Challenge 2025. Source code and model weights are shared on GitHub: https://github.com/Ahus-AIM/physionet-challenge-2025