{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/c43b02a7f4f8441b917138c0a9ba11fa\" frameborder=\"0\" width=\"1280\" height=\"960\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":960,"width":1280,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":960,"thumbnail_width":1280,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/c43b02a7f4f8441b917138c0a9ba11fa-78825d8065461f8a.gif","duration":145.328,"title":"Developing a R.A.G. Agent for Policy Questions with Accurate Citations 🚀","description":"Hi, I'm Nivejita, and in this video, I introduced my startup policy co-pilot, which is designed to answer policy questions with precise social citations using a R.A.G. agent. I've built it using technologies like Langchain, Chroma, and Fast API, and I demonstrated its functionality by running the server and querying an example about the seed fund scheme. The agent achieved about 80% accuracy on 10 FAQs with 100% correct citations. I also shared details about the four notebooks in my repository that support the project, including evaluation reports. I encourage you to check out my repository for more information and would love your feedback!"}