Evaluated machine learning adversarial attacks against proposed novel defenses for deep learning.Led a key submission process for an external organization to evaluate a researched adversarial defense.Helped develop a codebase for evaluating a novel defense against data poisoning.Isolated and helped fix a critical bug in a computer vision pipeline intended to port a published technique from TensorFlow to PyTorch for research oriented extensions. The bug fix came before an important deadline and improved performance as measured by mean average precision to match and even slightly exceed the original published performance of the TensorFlow implementation. The performance prior to the bug fix was significantly degraded and displayed key discrepancies relative to the original published performance. Gained familiarity with the adversarial machine learning literature such as evasion attacks and data poisoning through presenting various collections of papers for internal company meetings.Became acquainted with key papers in the deep learning object detection literature.