Work
My focus is AI-native products built around people: agentic and retrieval-based systems grounded in real user research, shaped by design, and supported by the infrastructure and enablement that get teams to actually adopt them.
These case studies are locked. If Carlos shared an access key with you, enter it below.
To request access, reach out to Carlos on LinkedIn.
Multi-agent productivity assistant
One trustworthy surface where product teams generate work artifacts, get research-backed answers, and reuse vetted prompts instead of hopping between tools. Shipped to production, and saved teams tens of thousands of hours.
Compliant retrieval assistant
Fast, traceable answers from vetted sources for researchers and product managers, with nothing to set up and no compliance risk. Owned end to end, delivered in three months, adopted by hundreds of people.
Marketplace for reusable AI skills
An experiment: a working marketplace for creating, sharing, and installing real agentic capabilities, not just prompts. Designed and built front to back to test whether non-technical people could author them. They could, and the learnings are now going into a product.
Enterprise AI adoption
Platform, curriculum, and coaching aimed at making AI habits stick at scale rather than just land. Lead author on a self-paced course taken by 20,000 people, and built the initiative's website.
Context strategy for AI-native teams
A framework for capturing how work actually gets done, so a team and its AI tools run on the same durable memory. Currently being adopted across teams.