Home / Resources / Guides

Guides

Practical guides for robotics data teams

Six guides covering scoping, evaluation, sourcing decisions, and the technical concepts behind embodied AI data.

All guides

Six guides, all under twelve minutes

Written by the team that runs the programs. No vendor fluff.

GUIDE 01

A four-step framework for going from “we need data” to a signed SOW. Volume targets, acceptance criteria, ramp planning.

12 min read • For technical buyers

GUIDE 02

The data class that separates humanoid policies that transfer from ones that don’t. Capture, retargeting, and validation.

10 min read • For ML leads

GUIDE 03

How to measure the sim-to-real gap, what “91% transfer” actually means, and the three pitfalls in synthetic pipelines.

11 min read • For sim engineers

GUIDE 04

The seven questions that separate operations-grade data partners from labeling marketplaces.

9 min read • For procurement

GUIDE 05

When each model fits, what each costs, and the IP and quality tradeoffs at each tier.

8 min read • For program leads

GUIDE 06

Why 30-minute episodes break generic pipelines, how we structure capture, and what failure-recovery data looks like.

11 min read • For research teams

What our partners say
The cross-embodiment guide is what I sent to my CTO when I was making the case for switching data partners. He read it in the car and called me back ten minutes later.
Wei Chen
VP Engineering, Holos Robotics

Want a guide on something else?

Tell us what your team is wrestling with. If three teams have the same question, we write the guide.