Figure
Helix AI Engineer, Agentic Systems
About this role
Figure AI seeks an AI engineer to develop autonomous agentic systems for humanoid robots that perceive raw sensory inputs, reason over memory, and execute complex tasks reliably in real-world environments. You'll design multimodal agents capable of operating continuously for extended periods, building the perception-reasoning-action loops and infrastructure that enable embodied AI.
What you'll do
- Design and deploy multimodal agents that operate autonomously from raw sensory inputs and execute multi-step tasks
- Implement episodic memory systems and long-horizon planning mechanisms for persistent reasoning
- Build reliable perception-reasoning-action loops with strong failure recovery and stability
- Apply reinforcement learning and post-training techniques to improve agent reasoning in real-world scenarios
- Design evaluation frameworks and benchmarks to measure robot reasoning, planning, and task success
- Develop scalable infrastructure for distributed model training and agent evaluation
What they're looking for
- Autonomous agent development
- Multimodal model training and fine-tuning
- Python and PyTorch
- Agent memory and planning systems
- Pixel-to-action reasoning
- Machine learning experimental design
- Software engineering and systems reliability
- Reinforcement learning and reward modeling
Benefits
- Competitive base salary: $150,000 - $350,000 annually
- Full-time position at AI robotics leader
- Work on cutting-edge embodied AI technology
- Additional compensation components available
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Figure
Figure develops advanced humanoid robots powered by AI technology. The company is hiring engineers across mechanical design, firmware development, manufacturing, quality assurance, and security to build and refine its autonomous robotic systems.
View all jobs at FigureLikely interview questions
- Can you walk us through a project where you built an autonomous agent that operated continuously for extended periods? What were the key challenges in maintaining reliability and handling failures?
- Describe your experience training or fine-tuning multimodal foundation models. What frameworks and techniques did you use, and how did you evaluate model performance?