Worked on the meta-pytorch/forge repository to establish foundational data modeling for conversational AI systems. Developed core data abstractions for prompts, completions, and episodes, creating a standardized structure for handling conversational data and model outputs. Leveraged Python and object-oriented programming principles to design scalable data models that support consistent dialogue workflows and facilitate future feature integration. The implementation focused on building a robust data layer, improving testability, and enabling easier integration of new components across the dialogue system. This work laid the groundwork for more advanced natural language processing capabilities by ensuring a reliable and extensible data pipeline within the project.