Products

Resources

Research

Featured · 31 July, 2026Assemble Bench: Benchmarking Robot Models on Contact-Rich Assembly

A NIST-taskboard assembly benchmark on Isaac Lab Arena and the DROID platform – 14 peg, gear, and nut tasks for training and evaluating VLAs, plus CG-DAgger improvement on π0.5.

Lukass Kellijs, The HUD Team
  • 20 January, 2026
    Engineering
    Building an RL Environment to Train Agents for Production DebuggingThe HUD Team
  • 1 October, 2025
    Enterprise
    Evaluating Agents on Financial Analyst Workflows (SheetBench)The HUD Team, Sepal AI
hudAnything you can simulate and grade, you can improve.
Product
Resources
Company
© 2026 Human Union Data, Inc.All rights reserved.