Taking AI to the playground: LinkedIn combines LLMs, LangChain and Jupyter Notebooks to improve prompt engineering
1 min read
Summary
Professional social network LinkedIn has developed a “collaborative prompt engineering playground” to allow both technical and non-technical staff to create prompts for generative AI (AI that uses large volumes of text data to create copy, instead of infinitely varying datasets of images or music, for example).
The system is based on a combination of large language models (LLMs), LangChain, and Jupyter Notebooks, and has already been used to develop an AI feature for its sales navigator product, reducing company research time from 2 hours to 5 minutes.
The combination of accessible interfaces and secure data sources allowed its users to experiment with different prompts and ideas for using gen AI, and to refine the outputs from those ideas quickly, in collaboration with domain experts.
The platform is deeply integrated with LinkedIn’s internal systems, so it’s unlikely to be offered as a standalone product, but its building blocks, including the LLM, LangChain, and Jupyter notebooks, are available to other organisations who want to develop a similar platform.