Worked on a major refactor of the LogicFormatter component in the finch-tensor-lite repository, focusing on enabling abstract format selection and parametric tensor creation. The approach involved introducing an abstract LogicFormatter class with a default implementation, allowing fill values to be passed as parameters during tensor creation for greater flexibility. Code was reorganized to better align with the project’s structure, and a tensor creator scaffold was added to support future extensibility. Addressed a missed reference to the logic formatter to improve stability and ran pre-commit checks to enhance CI readiness. Utilized Python, full stack development, and software architecture skills throughout.