tech
March 31, 2026
Nomadic raises $8.4 million to wrangle the data pouring off autonomous vehicles
The company turns footage from robots into structured, searchable datasets with a deep learning model.
TL;DR
- Companies building autonomous machines collect vast amounts of video data for training and evaluation.
- Organizing and cataloging this data, especially for rare edge cases, is a significant challenge.
- NomadicML offers a platform that uses vision language models to convert video footage into structured, searchable datasets.
- This allows for better fleet monitoring, creation of unique datasets for reinforcement learning, and faster iteration.
- NomadicML raised $8.4 million in seed funding, led by TQ Ventures.
- Customers like Zoox and Mitsubishi Electric are already using the platform.
- The platform is described as an 'agentic reasoning system' that can find specific information within footage.
- Future development includes tools for non-visual data like lidar and cross-modal sensor integration.
Continue reading the original article