Rapidata was founded in 2023 in Zurich, Switzerland, with CEO and founder Jason Corkill, who hit the data-annotation bottleneck firsthand while working in robotics and computer vision at ETH Zurich. The company targets one of AI's most persistent constraints: collecting large-scale, high-quality human feedback to train and evaluate models is slow, expensive and operationally painful, yet it is essential for reinforcement learning from human feedback (RLHF) and for honestly measuring model quality.
Rapidata's answer is a global human-feedback network designed for speed and scale. Rather than relying on small, slow panels of annotators, it taps very large pools of human responders, including through novel mechanisms like digital ad placements, to gather feedback in near real time. This lets AI teams run human evaluations and preference comparisons in days instead of months, dramatically shortening the iteration loop for improving models.
The platform serves both training and evaluation needs. On the training side, it supplies the preference data and human signals that power RLHF and alignment work. On the evaluation side, it lets teams benchmark model outputs, compare versions, and validate behavior against human judgment at a scale that traditional vendors struggle to match. By making human feedback fast and abundant, Rapidata aims to remove a key drag on AI development velocity.
In February 2026 Rapidata raised a 8.5 million dollar (7.2 million euro) seed round co-led by Canaan Partners and IA Ventures, with participation from Acequia Capital and BlueYard. The funding supports scaling its global feedback network to meet growing demand from AI labs and companies. As a Swiss infrastructure startup attacking the human-feedback bottleneck with a genuinely novel approach, Rapidata is a notable European entrant in the AI training-data and evaluation space.