A brain cancer or tumor is a condition occurred due to the expansion of unnatural nerve cells. Because tumors are rare and can take many different forms, it is challenging to estimate the survival rate of a patient who has been impacted. These tumors can be found using Magnetic Resonance (MRI) Images, which are crucial for locating the tumor placement; however, non-automatic recognition is a labor-intensive & difficult process that may yield inaccurate findings. Segmentation is also required to calculate the tumor's size and other prognostic parameters. Adopting computer-aided methods is crucial to assisting in overcoming these limitations. Various models of Deep learning are utilized in medical image analysis to detect brain tumor employing MRI images as artificial intelligence (AI) technology progresses. This paper presents a deep learning convolutional neural network fine-tuned Efficient Net baseline model and K-means clustering based segmentation are utilized to effectively detect and segment images of brain tumors, respectively. In order to boost the number of data samples for our suggested model's training, data augmentation techniques are used. The tumor is separated from the MRI images using K-means clustering. Brain tumor detection is carried out using fine-tuned Efficient-B0. The findings demonstrate that the state-of-the-art EfficientNet-B0 model, which has been suggested and fine-tuned, has obtained excellent classification accuracy, precision, and recall values, with final accuracy of 98.66% for overall segmentation and detection.