Worked on the pytorch/executorch repository to deliver an internal QuantizedLinearWeightBase implementation, replacing the previous reliance on torchao.quantization.subclass. This approach decoupled Executorch from an external quantization dependency, reducing maintenance complexity and external risk while preserving existing API behavior. To ensure backward compatibility during the migration, a legacy placeholder subclass was introduced, allowing seamless transition for users. The work was implemented using Python and leveraged deep learning and machine learning expertise, particularly with PyTorch. No major bugs were addressed during this period, with the primary focus on feature development and architectural improvements to support long-term maintainability and stability.