# floatingpoint

> floatingpoint builds supervised datasets, evaluations, and visual task environments for AI teams.

## Company Overview

floatingpoint is a data research lab focused on artificial visual intelligence. We build off-the-shelf and custom data products for teams training, testing, and shipping vision-capable models. Our work spans document intelligence, general vision, computer use, visual software engineering, spatial reasoning, and long-horizon visual tasks.

## Core Products

- Supervised Datasets: High-quality supervised data for layout analysis, table structure, form understanding, structured extraction, schematic analysis, object and instance counting, and other visual primitives.
- Long-horizon Tasks: Hand-crafted environments and verifiers for document reasoning, computer use, visual software engineering, game design, and spatial reasoning.
- Benchmarking: Public and private evaluations for measuring visual capability gains in model systems.

## Bespoke Services

floatingpoint works with ML teams as a research partner to build custom supervised datasets. Engagements usually cover schema design, distribution design, annotation guides, QA tooling, ML-assisted review, golden-set delivery, scaled production, and ongoing evaluation.

## Process

1. Alignment: Co-design the schema, distribution, annotation guide, and edge cases.
2. Tooling: Build custom QA, annotation, and ML-assist systems.
3. Golden set: Deliver a small dataset so teams can validate the recipe.
4. Scale: Ramp production volume against the agreed distribution.
5. Delivery: Send data as it clears QA, with evaluation results and learnings.

## Contact

- Website: https://floatingpoint.ai
- Schedule a call: https://cal.com/floatingpoint
- Email: contact@floatingpoint.ai
- Twitter/X: https://twitter.com/floatingpointai

## Pages

- Homepage: https://floatingpoint.ai