Developed and integrated an image-based text generation feature within the langchain-ai/langchain repository, expanding HuggingFacePipeline to support multimodal tasks by enabling extraction of generated text from images. Focused on backend development and API integration using Python, the work ensured seamless interoperability with LangChain for enhanced business value. Subsequently, addressed a bug in the liguodongiot/transformers repository, refining the ImageTextToTextPipeline to correctly process user-defined generation configurations. This fix improved the flexibility and reliability of image-to-text workflows, leveraging deep learning and NLP techniques. The contributions demonstrated depth in full stack development and a methodical approach to both feature delivery and maintenance.