Abhimanyu implemented a neural network dropout customization API for the sktime/sktime repository, focusing on enhancing regularization control in deep learning models. Using Python and Keras, he enabled users to specify dropout rates for MLPNetwork and RNNNetwork either as a uniform float or a per-layer tuple, supporting both flexibility and backward compatibility. The solution included validation to ensure tuple lengths matched network layers and preserved existing defaults to prevent regressions. By addressing issue #9103, Abhimanyu improved API stability and usability, facilitating safer experimentation and onboarding for time-series modeling. The work demonstrated thoughtful design and careful integration with existing code.