Most enterprise AI platforms today still force domain experts to learn how to prompt. At Cogzia, we are flipping the paradigm: the platform learns from the expert.
‍
📍 Building Autonomous Infrastructure at Cogzia As Co-Founder and CTO of Cogzia, I architect multi-agent orchestration frameworks for highly regulated industries like oil & gas, life sciences, and fintech. My focus is on building AI that doesn't wait to be asked.
‍
Using our "Council of Experts" architecture and proactive agentic layers, we build a persistent, compounding understanding of a team’s data and workflows. We transform messy, unstructured enterprise data into hot-loadable MCP tooling and highly structured knowledge artifacts. Our system acts like a senior business analyst—surveying schemas, surfacing insights, and contextualizing anomalies before a single prompt is typed. We are building systems that stop acting like passive software tools and start acting like proactive team members.
‍
📍 A Decade of Pioneering Enterprise AI The architecture at Cogzia is the culmination of a decade spent building production-grade ML systems. Previously, as Co-Founder and CTO of Huma.AI (Human + AI), I realized early that the true enterprise moat wasn't in training custom foundation models, but in orchestration.
‍
Years before the industry had a name for it, I built one of the earliest production Retrieval-Augmented Generation (RAG) systems in the enterprise. This infrastructure allowed biotech researchers to query millions of scientific publications in their own language. We partnered with OpenAI, AWS, and Google to prove a core thesis: when you build AI around the expert instead of the model, you get systems people actually trust.
‍
📍 The Foundation My conviction that software should adapt to the human—not the other way around—started on day one of my career. I started out doing product demos shoulder-to-shoulder with Steve Jobs at Claris. I was honestly too young to be intimidated, so we would just argue over the finer points of the UI. That experience burned a core philosophy into me that I still apply to complex ML pipelines today: the technology only matters if the human interacting with it can easily wield it.
‍
Always happy to connect with other founders, deep tech investors, and AI pragmaticists.

‍