During April 2025, Thang Nguyen enhanced the NVIDIA/bionemo-framework by introducing baseline performance metrics for the ESM2 model within the continuous integration pipeline. Leveraging skills in benchmarking and CI/CD, Thang implemented YAML-based configurations to automatically track key metrics such as consumed samples, validation loss, reduced training loss, and TFLOPS per second per GPU. This addition enabled automated validation of execution metrics, supporting data-driven optimization and faster regression detection across model runs. The work focused on improving CI observability and benchmarking depth, laying the groundwork for more reliable performance monitoring and streamlined model evaluation within the repository’s development workflow.