Due to their impressive natural language capabilities, large language models (LLMs) are paving the way for fast prototyping of NLP applications in any domain. In this talk, I showcase how to build a structured NLP pipeline to mine clinical trials, using the open-source NLP toolbox spaCy and its recent extension ‘spacy-llm’. Moving beyond a fast prototype, she’ll discuss how to manage different performance features such as accuracy, speed, memory usage, reliability, maintainability and customizability, and offer pragmatic solutions to make the pipeline more reliable and cost efficient.

→  Venue: AstraZeneca NLP Community of Practice (external speaker, online)

→  Slides: SpeakerDeck

→  Repo: Github