Helix AI Engineer, Agentic Systems
Core
Building autonomous multimodal reasoning systems (agents) that perceive via pixels, maintain episodic memory, and execute complex tasks in real-world environments for humanoid robots.
Role type
Senior IC machine-learning engineer (agentic systems)
Builds
Autonomous embodied AI agents and the underlying infrastructure for robot autonomy
Domain
Robotics + Embodied AI
Deliverable
production ML models
Required skills
Multimodal agent architecture, episodic memory systems, long-horizon planning, reinforcement learning, reward modeling, Python, PyTorch, distributed training, experimental design
Preferred skills
Embodied AI/robotics learning, multimodal foundation models, agentic AI systems, large-scale distributed training, publication record in ML/robotics
Technologies
PyTorch, Python
Responsibilities
Design, train, and deploy multimodal agents for hours-to-days autonomous operation; Build agents reasoning from raw sensory inputs to structured actions; Implement episodic memory systems for persistent state and retrieval; Develop planning, reasoning, and tool-use mechanisms; Build reliable perception → reasoning → action loops with failure recovery; Design evaluation harnesses and benchmarks for agent reliability; Run data studies across the training lifecycle; Apply RL and post-training techniques to improve agent performance; Build infrastructure for scalable model training and experimentation; Integrate agent models into the full humanoid autonomy stack
Seniority
Senior, hands-on IC
