Designed a robust back-end API utilizing Django and Celery, optimizing service performance by leveraging OpenAI API. Continuously performed prompt engineering to enhance model effectiveness.   Programmed a reusable module adopted by various teams and projects, enabling response streaming through WebSocket technology.   Trained machine learning models and maintained data collection and processing system used for  fine-tuning a large language model. Constructed an API and client interface for internal usage.

더보기

1. I made significant performance improvements to the service by transitioning from Parquet  les to Arrow  les for compression and switching from Pandas to Polars for  le processing. These changes resulted in a performance boost of over 3 times. 2. I also played a key role in expanding the capabilities of the backend API (FastAPI) and Python client by implementing a new feature that enabled data export. This enhancement provided users with an additional way to interact with the service. 3. I contributed to data processing for a machine learning model, utilizing PyTorch. By providing a clean and reliable source of data, I supported the development of a robust language model. 4. I automated the process of updating the Python package documentation by integrating Git and Sphinx. I also implemented user authentication and authorization to the documentation system using JWT, enhancing security and user access control.

더보기