•Designed, built, and shipped systems for evaluation of large language models as part of the Amazon Bedrock product, allowing thousands of customers to quickly and easily assess modern LLMs for accuracy, toxicity, and other dimensions.•Drove development of data annotation tools for demonstration data collection, preference ranking data, red teaming, and other generative AI use cases, leading to new multi-million dollar business opportunities.•Deployed and integrated a novel PyTorch model for bounding box annotation, allowing customers to reduce their bounding box annotation time by 39%.•Founding team member of the SageMaker Ground Truth service, where I built distributed systems for running active learning models, allowing customers to save 70% on their data annotation costs.•Tech lead with responsibility for steering weekly engineering operations meetings.•Comfortable working across the software lifecycle, including designing, shipping, and operating internet-scale web services.•Fluent in core AWS technologies, including S3, EC2, AWS Lambda, API Gateway, SNS, SQS, and others, along with Java, Python, and Docker.

ATA provides aerospace engineering services including testing and analysis for clients including NASA. During my tenure I developed expertise in implementing numerical methods using Python to address quantitative, computational research problems.I served as principal investigator on a $1M DoD software research project to develop computational tools that accelerate the development of composite materials, which often takes years and millions of dollars. This project required me to build and train neural networks using Python, TensorFlow, and Keras to predict the mechanical behavior of composite materials. I implemented Bayesian statistical models in Python, collaborative filtering, and non-negative matrix factorization algorithms to predict unknown composite material modeling parameters. I also optimized carbon fiber composite ply orientations using genetic algorithms and particle swarm optimization in Python.While at ATA I written or co-authored over 20 proposals with a 75% win rate, collectively worth $5M. I also supervised a junior software developer and interviewed over 20 engineering candidates as a member of the company interview team.

As a Research Assistant I conducted large-scale numerical simulations on state-of-the-art supercomputers and post-processed hundreds of gigabytes of data and automated data extraction from a proprietary database using Python scripting tools and the database API. My research was undertaken as part of a thesis-based master’s degree in mechanical engineering, which also led to the publication of two refereed journal articles.

As part of my co-op I performed static and dynamic structural analysis of mechanical and aerospace systems.