Authors - Aditya Gujrathi, Sarvesh Morey, Tushar Mahajan, Ganesh Masute Dhanraj Jadhav, Avinash Golande, Kirti Deshpande Abstract - This research explores the latest developments in deep learning architectures to achieve higher accuracy in face recognition. We examine the efficacy of well-known models such as VGGFace and Face Net, investigating their capacity to extract unique facial traits that distinguish people under difficult situations. In order to enhance model generalization, the study explores the effects of data augmentation approaches, which create variations of preexisting images in order to artificially extend training datasets. We also go over methods for reducing bias in training data, which is an important part of making sure the model is equitable across a variety of demographics. The study also looks at how the model's performance is affected by various loss functions, which direct the learning process. Lastly, we suggest some directions for additional study to improve the precision and resilience of deep face recognition systems.