Over the years, information extraction has undergone numerous significant changes. This also holds true for the medical field. Medical transcripts are valuable resources filled with information that can help medical professionals and patients with insights. However, extracting this data manually is not efficient. Automating or minimalizing the process is required to perform efficiently. We aim to achieve No Code data extraction from medical transcripts to achieve inclusivity. To achieve this, we fine-tune the model using curated datasets. LLMs are powerful learners, but lack specific knowledge. By feeding them medical data like research papers and clinical documents, we adjust their internal understanding towards the medical domain. This lets them recognize medical terms, grasp relationships between diseases and treatments, and understand the context of medical queries. So, when we ask a question, the fine-tuned LLM can search through the data it learned to give us an accurate and relevant answer [5, 6]. We evaluate the model’s performance using RAG metrics for data retrieval. Compared to Mixtral 7B, our proposed model shows 8% improvement in MMLU performance. Translation being the only weak section of our proposed model as it is trained for data retrieval and not translation. The fine-tuning process is carried out with fixed parameters listed in the PEFT section. Methodology of training is PEFT and for learning, reinforcement learning with a learning rate of 1e−4. This value is experimental and the relation between learning rate and results vary due to difference in nature and quality of the training data.