Brook Riggio is a software engineer, technology educator, and co-founder of Code Fellows, the Seattle coding academy that trained 3,338 graduates. His work spans accessible developer education, the human consequences of AI-assisted programming, and dependable architectures for user-facing AI agents.
From coding classrooms to AI applications
Riggio studied computer engineering at Harvey Mudd College and ran a software consultancy before helping launch Code Fellows in 2013. As a founding instructor and vice president of education, he taught Ruby on Rails, JavaScript, computer science, and web development while shaping an immersive curriculum focused on employment and accessibility. His published writing addressed beginning programmers’ practical challenges, from choosing a language to preparing for technical interviews.
He continued teaching foundational concepts beyond the academy, leading a 2024 CascadiaJS workshop on JavaScript trees and hash tables that welcomed developers without computer-science degrees. Before Code Fellows closed its classrooms, Riggio helped incorporate AI coding tools, AI-powered APIs, retrieval-augmented generation, and agents into its curriculum and operations; his reflection on its closure documented its graduate count and educational mission.
- Human judgment in AI-assisted development: Riggio frames a developer’s value around understanding the problem, the audience, and the context surrounding generated code. His GitHub Universe discussion asks engineers to articulate what remains distinctly human as implementation becomes easier to automate.
- Serverless full-stack AI architecture: His open-source serverless-agents repository combines Next.js, Python-based OpenAI Agents SDK workflows, Inngest, and Vercel into a deployable reference application. A research agent gathers information on user-selected topics, a second agent formats a newsletter, and the frontend retrieves the finished result from storage.
- Event-driven orchestration for dependable agents: Riggio structures agent execution as observable steps with retries and typed handoffs. His AI Engineer World’s Fair demonstration highlights an easily missed operational hazard: code outside explicit workflow steps can run repeatedly when orchestration resumes.
