Connecting human intent to agent execution—the research infrastructure that makes autonomous software development actually serve people.
AI coding assistants are transforming how software gets built. But speed without direction just means building the wrong thing faster.
Deutero encodes gold-standard qualitative research methods from social science—interviewing and thematic analysis—into cybernetic form. We teach AI agents how to learn about humans, translating techniques honed over decades into systematic processes that reveal what users actually need.
Named after Gregory Bateson's concept of "deutero-learning"—learning how to learn—Deutero bridges the gap between human intent and agent execution. As autonomous systems proliferate, we're building the human opinion discovery layer for the agent economy: ensuring that autonomous development serves human purposes, not just technical metrics.
Founder & CEO
PhD (NYU) applying Science & Technology Studies and sociology of finance to Silicon Valley—examining the epistemology of Lean Startup methodology and how tech companies come to "know" their users.
Postdoc (UBA) researching the use of AI agents to conduct social research via qualitative interviews—the foundational work that became Deutero's core technology.
Developed a formal information-theoretic model of qualitative interviewing and optimization strategies for interviewing agents. These methods are now applied in Deutero's product to deliver research-grade insights at the speed AI development demands.
