Conference affiliation: Head of Developer Relations · Arize · 2026
On this page
Laurie Voss is the co-founder and founding chief technology officer of npm, Inc. and the head of developer relations at Arize AI. After helping build essential infrastructure for the JavaScript ecosystem, he now concentrates on making AI agents observable, testable, and reliable.
From the early web to AI infrastructure
A web developer since 1996, Voss spent four years at Yahoo, where he worked with future npm co-founder Isaac Schlueter. He co-founded the analytics company awe.sm in 2010 and reunited with Schlueter to establish npm, Inc. in 2014. As the company’s founding CTO, he brought experience in large-scale web services, databases, and analytics to the infrastructure surrounding JavaScript’s package manager.
Voss has also taught Stuff Everybody Knows Except You, an introduction to professional web development, since 2014. He subsequently worked at Netlify as a data analyst and evangelist before becoming vice president of developer relations at LlamaIndex, where he focused on document processing, retrieval, and production agent architecture. He later joined Arize AI to lead developer relations, shifting his attention toward agent evaluation and observability.
How Voss thinks about dependable AI
Agents should solve concrete software problems. Voss favors systems that turn messy documents into actionable outputs—interpreting contracts, processing invoices, applying regulations, or producing financial reports. His approach to production agent design incorporates routing, parallel execution, orchestrator-worker systems, and evaluator-optimizer loops, while locating their value inside existing applications rather than standalone chatbots.
Retrieval-augmented generation remains practical infrastructure. Selecting relevant context is typically faster and cheaper than processing an entire corpus. Agents can improve retrieval by decomposing complicated questions, repeating weak searches, and checking whether their answers actually address the task.
Agent evaluation starts with understanding failure. Voss combines execution traces, inexpensive deterministic checks, human-labeled examples, and LLM-as-a-judge rubrics. His hands-on evaluation workshop demonstrates why a superficially plausible financial report can still concern the wrong company, omit an investment recommendation, or be delivered incorrectly.
Evaluators need their own safeguards. Generic correctness checks can fail when a judging model lacks current information; a faithfulness evaluation instead tests whether an answer reflects the research supplied. Voss recommends checking automated judgments against labeled examples, watching for bias, and avoiding tests so prescriptive that they reject better solutions.
Capability and regression evaluations serve different purposes. Capability tests define behaviors an agent has yet to master; regression tests preserve behaviors that already work. Voss prioritizes better data before prompt changes, model selection, or parameter tuning, and regards fully automated optimization as promising but immature.
