Supervised learning tasks particularly in the domain of Named Entity recognition are computationally costly and require significant effort, particularly in the biomedical domain where large amounts of annotated data may not always be available. To address these challenges the proposed research study adapts and seeks to work on Zero-shot and few-shot mechanisms for Named Entity Recognition. In the proposed research study, a fine-tuned BioBERT-based model has been used to implement zero-shot and few-shot models focused on accurately detecting the disease class. The approach followed achieves a high F1 score of 78.89% demonstrating the model’s effectiveness in accurately detecting the disease class. This high F1 score underscores the potential of zero-shot and few-shot mechanisms in Named Entity Recognition, particularly in the biomedical domain where annotated data is scarce.