Contributed token-level classification support to the NVIDIA-NeMo/Automodel repository, enabling tasks such as named entity recognition within the automodel framework. Developed a new model class and integrated necessary imports to facilitate token-level predictions, enhancing the flexibility of downstream NLP pipelines. Focused on improving code quality by addressing import issues and performing linting cleanups, which stabilized the build and streamlined future contributions. Leveraged deep learning and natural language processing expertise using Python to ensure maintainability and extensibility for token-classification workflows. The work reduced integration effort for users and accelerated deployment of token classification models within the existing NeMo automodel ecosystem.