Raven
AI model give response based on localization audio prompt and then review it to see the accuracy of the response. The prompt writer then give a review based on the model response if there any problem occurs. Prompt type need to be appropriate with task given either it is closed Q&A, open Q&A, role playing, creative writing, summarization or conversation. this project required LLM, Benchmark, QA/ Testing and Failure-Mode Analysis. This project required a space with natural background noice like people talking or vehicle noise. I succesfully completed the task although i had to find a place where it fit the task.