David Petersen

About David Petersen

Staff Software Engineer focused on applied AI and real-world applications of LLMs. I\'m passionate about thinking through how to apply large language models to solve practical problems—particularly in domains like customer support, content understanding, and infrastructure. With over 10 years of software engineering experience, I bring a pragmatic approach to building AI-powered systems that need to handle unbounded input and output spaces while maintaining reliability and security. Always exploring the intersection of cutting-edge AI research and production-grade engineering.

Experience

1. Staff Software Engineer

   Meta

   Nov 2024 — Present · Bellevue, WA, US

   Focused entirely on applied AI, building LLM-powered internal agents that operate across a wide input space and handle far reaching potential outputs. This means designing systems that can understand any customer question—regardless of how it\'s framed—and generate accurate, contextually appropriate responses without relying on predefined templates or rigid decision trees.Key challenges include managing hallucination risk at scale, building robust evaluation pipelines for open-ended generation, and ensuring agent responses remain safe and policy-compliant. This work sits at the intersection of production engineering and applied machine learning, requiring deep thinking about how to make LLMs reliable in high-stakes, real-world settings.Started by working in privacy infrastructure.

Education

• Oregon State University

  Computer Science

  2007 — 2011