Aaron Beppu
About Aaron Beppu
Aaron Beppu is a Staff ML Engineer at Hyperscience with extensive experience in software engineering and machine learning, having previously worked at Sift Science, Prismatic, Etsy, and A9.com.
Title and Role at Hyperscience
Aaron Beppu holds the position of Staff ML Engineer at Hyperscience. He has been contributing to the company since 2021, working remotely from San Francisco, California. His role focuses on developing machine learning models and tools tailored to customer-specific business needs. Aaron's expertise in optimizing model training processes and enhancing document annotation workflows is integral to his contributions at Hyperscience.
Previous Role at Sift Science
Before joining Hyperscience, Aaron Beppu served as a Principal Software Engineer at Sift Science for eight years, from 2013 to 2021. During his tenure, he played a significant role in enhancing the company's machine learning capabilities, which included optimizing training bottlenecks and implementing incremental training methods. His work contributed to Sift Science’s ability to offer sophisticated fraud detection solutions.
Experience at Prismatic, Etsy, and A9.com
Aaron Beppu's professional journey includes short stints and formative roles across various companies. In 2013, he worked as a Software Engineer at Prismatic for five months, located in the San Francisco Bay Area. From 2011 to 2013, he was a Software Engineer at Etsy, focusing on engineering tasks that supported the platform's robust e-commerce operations. Prior to that, Aaron worked as a Software Development Engineer at A9.com from 2008 to 2010, contributing to the company's search technology initiatives.
Educational Background
Aaron Beppu studied Cognitive Science at the University of California, Berkeley, earning his Bachelor of Arts degree in 2008. His three-year tenure at the university laid the foundational knowledge in cognition and computational sciences, which later translated into his ability to develop and enhance machine learning models and software engineering solutions throughout his career.
Technical Contributions and Innovations
Throughout his career, Aaron Beppu has made several technical contributions and innovations. He developed models capable of performing effectively with small, custom datasets and limited hardware, addressing specific business needs of customers. He implemented tools to improve document annotation processes, flagged anomalous annotations, and optimized training bottlenecks to enhance model performance. Additionally, Aaron introduced tools for real-time observability metrics and profiling in ML libraries, incorporating Ray parallelization to elevate the efficiency of ML pipelines.