At Aledade, Inc., I contributed to building robust backend systems and data architectures that significantly improved operational efficiency. My role involved designing event-driven solutions and APIs for real-time data processing, while also leading automation initiatives to enhance team workflows. The innovative technologies I employed included Python, AWS, and Kafka, which collectively drove substantial performance improvements.

Working on the project 'Semantic Segmentation using Meta-Learning for Medical Images' under Dr. Cauligi Raghavendra and Dr. Ajitesh Srivastava at USC.

• Developed and implemented a novel meta-learning framework called Amphibian as part of the Medical Decathlon Challenge.

• Focused on improving medical image segmentation for computer-aided diagnosis systems.

Successfully addressed the limitations of state-of-the-art (SOTA) models when dealing with unseen tasks and limited annotated data.

• Leveraged dynamically penalized weighted task subsampling to reduce sampling bias across organs during the meta-training process.

• Achieved remarkable segmentation scores across various unseen organs such as the spleen, prostate, and heart.

• Demonstrated the effectiveness of the Amphibian framework in the context of the Medical Decathlon Challenge.

Maintaining the USC room finder website with the latest-in-class photographs, 360-degree view, and class amenities.

Part of the team responsible for modernizing technology and improving the learning environment for the central IT services that support the academic institutions, hospitals, research facilities, and administrative divisions of USC.

• Developed a deep learning approach to denoise human and mouse brain MRI images along with the team using auto-encoder architecture.

• Obtained a cleaner image compared to traditional image processing approach employed in the BrainSuite software.