Helix Ai Engineer, Pretraining

Humanoids Daily

San Jose, CA, United States
On-site
Large-scale foundation models
Multimodal data pretraining
Transformer architectures
Figure is developing autonomous general-purpose humanoid robots with embodied AI systems that can perceive, reason, and act in the real world

Job Summary

  • Figure is developing autonomous general-purpose humanoid robots with embodied AI systems that can perceive, reason, and act in the real world.
  • The Helix team is responsible for developing core AI systems, focusing on pretraining large-scale foundation models across multimodal data for enhanced generalization, reasoning, and adaptability.
  • This role requires 5 days/week in-office collaboration in San Jose, CA, and involves designing and training models, optimizing training pipelines, and collaborating with various AI teams.

Matching Summary

Figure is developing autonomous general-purpose humanoid robots with embodied AI systems that can perceive, reason, and act in the real world.

Skills & Requirements

Must-have

  • Large-scale foundation models
  • Multimodal data pretraining
  • Transformer architectures
  • Large-scale distributed training
  • Python and PyTorch proficiency
  • Scalable, reliable systems

Nice-to-have

  • Frontier foundation model experience
  • Multimodal pretraining expertise
  • Scaling laws and dataset mixture
  • Post-training techniques
  • Embodied AI and robotics familiarity
  • Publication record in ML/NLP/Multimodal AI

Key Requirements

  • Experience training large-scale foundation models
  • Experience with LLM or multimodal pretraining
  • Strong understanding of transformers
  • Experience with large-scale distributed training
  • Proficiency in Python and PyTorch
  • Solid software engineering skills
  • Ability to drive ambiguous technical problems

Work Rights

Not specified

Tailored Resume

Cover Letter