PulaCloud was a venture built around human computation: using distributed human workers to solve problems computers could not handle well, while creating employment in developing countries.

The core work was building the platform layer that makes that kind of labor usable and trustworthy: task definition, worker onboarding, trust and quality scoring, workflow design, and the measurement of outcomes. A major motivation was applying the approach to semantic annotation and health-related visual data, where “ground truth” is expensive and computers alone often fail.

PulaCloud did not scale commercially, but the lessons carried forward: how to design trust systems, how to coordinate distributed work, and how to turn messy real-world activity into a repeatable operating system.