Spiral builds data infrastructure for AI teams. Its principal product is a fast analytical database purpose-built for multimodal, multi-rate data streams, operating on top of the open source Vortex file format. The system is designed to provide fast queries for data exploration and GPU-saturating throughput for model training, addressing the data-loading bottlenecks and idle GPU time that slow down research and engineering work.
A second product is a dataset creation and versioning platform, which lets teams turn raw robot data into exact, tested and versioned datasets for training and evaluation. Each iteration is reversible and shareable, so that work done on curation is not lost between runs.
The company's technology spans analytical databases, multimodal data infrastructure, I/O optimisation, dataset versioning and curation, and pipelines for time-series, video and sensor data. It is used across verticals including weather and climate, financial data, genomics, point clouds, video, images and robotics, and it maintains the open source Vortex file format.
Spiral has a presence in New York City, San Francisco and London.
