Transformer_TalentSYSTEM LIVE

/roles — ROLE_157

Staff Machine Learning Engineer, VLA Training

Early-stage robotics company automating American defense production lines with vision-language-action models

The role

COMP
$250K - $300K
EQUITY
Competitive equity
LOCATION
New York
WORKPLACE
On-site
EXPERIENCE
6 - 15 years
VISA
None
STACK
PyTorch
INDUSTRY
AI, Defense, Hardware, Manufacturing, Robotics, Software Development

The company

Early-stage company building AI and robotics systems for American defense manufacturers, letting robot arms and other factory machines run around the clock to raise output without extra staff.

STAGE
growth-stage
FUNDING
$30M+ raised
TEAM
~10 people
FOUNDED
2025

JD — the work

About the role

A hands-on applied ML role owning how vision-language-action models are trained, tested, and deployed on autonomous robots in American defense manufacturing. The team of about ten already earns meaningful revenue from a major defense contractor and is growing fast. This is applied work rather than pure research: the goal is models that run more reliably and faster in real deployments, and you would create systems from a blank slate with hardware on hand for quick real-world testing. The role is on-site in New York.

What you'll do

  • Improve VLA models through repeated training runs until they meet a live customer's reliability and latency goals
  • Run experiments on datasets, training recipes, and hyperparameters to learn which changes lift success rates
  • Stand up and grow multi-GPU distributed training infrastructure so experiments run quickly
  • Own the model lifecycle end to end: dataset construction, training, evaluation, and validation on robots
  • Expand experiment throughput across more workflows as operations grow

What they're looking for

  • 6+ years of machine learning engineering experience
  • Deep PyTorch skills and experience training large models on multi-GPU clusters
  • A preference for applied, production ML over pure research
  • Motivation to see your models drive physical machines outside the lab
APPLY FOR THIS ROLE →All open rolesOne application covers up to 3 roles.