Software Engineer Intern
Collaborated with research group on the development of a black hole merger detection model. Optimized model training to maintain accuracy despite increased parallelization. Improved training time by fifty percent, allowing the model to train on twice as much data in the same time frame. Enabled the detection of mergers across a wider signal manifold. Worked with TensorFlow and Keras in a high-performance computing environment, taking advantage of the Horovod framework for distributed training. Presented research poster to professors and students. Continued to work full time through the summer and part time during the school year to enhance model training efficiency.