Shiro enables teams to do rapid prototyping of prompts. You can perform a quantitative and qualitative analysis of multiple prompts, parameters, models, and even model providers across a variety of test cases.
So it makes it easy to quickly prototype, deploy, version, and monitor LLM calls. You can create complex prompts with variables, tags, and filters using the Liquid templating engine.
Prompts can then be deployed for use in production environments and accessed via our API, either to execute the call to the LLM or to access the prompt value. This allows developers to deploy a prompt while still optionally enabling technical or non-technical users to modify the prompt without updating code.
Last year I applied for a "Senior Prompt Engineer" position with Khan Academy. I was rejected, so I used my demo project from the application to launch Shiro. This is my story.