Helix AI Engineer, Video Pretraining (Video Foundation Models) at Figure AI (San Jose)

September 14, 2026
US$200,000 - US$400,000 / year
Application ends: October 14, 2026
Apply Now

Job Description

Figure is hiring a Helix AI Engineer to lead video pretraining: training large video foundation models on internet-scale video and robot-collected data, so that its humanoids learn motion, interaction and time from raw video. The pretrained models feed perception, prediction and embodied reasoning downstream.

At a glance

  • Team: Helix AI
  • Location: San Jose, California (five days a week in the office)
  • Pay: $200,000 – $400,000 a year, as listed
  • Apply: open until filled (re-checked 14 October 2026)

What you would do

  • Design and train large video foundation models on internet video and robot data
  • Develop pretraining strategies that capture temporal dynamics, motion and object interaction
  • Explore transformer- and diffusion-based architectures for video understanding and generation
  • Build efficient data pipelines and large-scale distributed training
  • Create benchmarks for temporal understanding, and integrate models with generative, agent and robot learning teams

Why this role matters

Robots cannot easily collect the billions of examples that language models learned from, but the world has enormous amounts of video of people doing things. Learning from that video is one of the most promising ways to give robots common sense about motion and objects, and Figure is betting on it as a core part of its AI stack.

What Figure AI is looking for

  • Experience training large-scale models, ideally on video
  • Knowledge of transformer and diffusion architectures
  • Distributed training and data pipeline skills

Nice to have

  • Experience applying video models to robotics or embodied AI

Pay and location

Figure lists $200,000 to $400,000 a year for this role.

How to apply

Apply through the official Figure AI job posting. Figure AI’s posting does not give a closing date, so the role is open until filled; ResearchJobs.in will re-check this listing on 14 October 2026. Check the posting for location and work-authorisation details before you apply.

See all our artificial intelligence jobs, or browse more research jobs on ResearchJobs.in.

Hiring institution: Figure AI

Official advertisement: job-boards.greenhouse.io

How to prepare for this application

  • Scale experience: show the largest models and datasets you trained.
  • Video specifics: temporal modelling, tokenisation or video diffusion work is directly relevant.
  • Efficiency: data-loading and training efficiency matter at video scale.
  • Link to robots: say how video pretraining could help control.

About Figure AI

Figure builds general-purpose humanoid robots, including Figure 03, for commercial work and the home. Its Helix team develops the AI that lets the robots see, reason and act. Figure is based in San Jose, California, and asks for five days a week in the office.

We send one confirmation email first. Every alert has an unsubscribe link.