طراحی زنجیره RAG برای درک متن (RAG Pipeline for Text Understanding)

توضیح این پروژه به فارسی با وجود اصطلاحات مختلف که معادل مناسبی ندارند کمی دشوار است و به همین دلیل من هدف پروژه را به انگلیسی توضیح می دهم.Our goal is to develop a retrieval augmented generation (RAG) pipeline for chat-based question and answerig based on an input text. The RAG pipeline should be optimized to give accurare results solely based on the text, with adjustable parameters such as the length of the response, etc. The pipeline should be eventually available through a simple front-end for a user. We need a skilled engineer in the field of AI, familiar with large/small language models who has experience in working with natural language data, text embedding models and open/closed source language models. The engineer should be able to offer us various options on how to implement the RAG pipeline based on various criteria.  Upon successful completion of this project, we are looking for solutions for more advanced tasks such as fine-tuning of SLMs/LLMs based on given input texts for a chat model with better performance. This task involves data preparation (labled questions and answers or few shot learning) as well as finding the right language model for fine-tuning. In case of an agreement, we would like to continue the work with the person taking over the first project in the follow-up projects with competitive payments through the ParsCoders platform.