Jobs at Figure.ai

Annotate and label images and video from humanoid robots using proprietary tools, collaborate with ML engineers on guidelines, flag edge cases, maintain quality and throughput, and provide feedback to improve workflows.

Own company-wide endpoint engineering across macOS, Windows, and Ubuntu fleets, including zero-touch provisioning, security hardening, patch compliance, EDR coverage, remote recovery, and low-connectivity device programs. Lead Okta SSO, RBAC, and identity lifecycle automation integrated with business SaaS platforms. Build scripts and integrations to eliminate manual IT work, manage audits and change processes, document standards, resolve escalations, and mentor IT operations staff.

Own end-to-end People Operations, including employee lifecycle administration, HR inquiries, leaves, accommodations, payroll, benefits, timekeeping compliance, and multi-state workforce support. Administer HR systems, workflows, permissions, integrations, and data governance while partnering on compensation architecture, knowledge management, compliance, vendor relationships, and cross-functional People projects. The role requires strong California employment law, FLSA, benefits, payroll, and contingent workforce expertise in a five-day, in-office environment.

The Deployment Engineer will support engineering during development and deployment at customer sites, develop testing plans, execute tests, analyze data, and communicate feedback.

Design and build test fixtures for humanoid robots, ensuring integration and validation of hardware through hands-on testing with prototype equipment.

The Robotic Safety Engineer will define safety strategies and ensure compliance for Figure's humanoid robot, focusing on obtaining safety certifications and supporting safety-critical functions.

Develop, train, deploy, and evaluate reinforcement learning algorithms for the whole body control of humanoid robots, focusing on performance optimization and robustness.

Lead the development of large-scale video foundation models for humanoid autonomy, focusing on training strategies and model evaluation for real-world applications.

Develop and enhance visuomotor manipulation policies for humanoid robots, focusing on real-world deployment and integration with other systems.

Develop learning systems using reinforcement learning for autonomous robots, focusing on policy optimization and real-world application.

The Helix AI Engineer, Modeling will design core model architectures for AI systems, focusing on multimodal inputs and improving model capabilities for humanoid robotics.

The Helix AI Engineer will design, train, and deploy large-scale generative models, focusing on diffusion modeling for robots, while enhancing robot autonomy through perception and world modeling.

As a Helix AI Engineer focused on pretraining, you will design and train foundation models using multimodal data, improve generalization strategies, and optimize distributed training pipelines.

Design, implement, and optimize the graphics rendering stack and UI framework for robot head displays. Build Qt/QML applications on embedded Linux, author shaders, implement Vulkan/OpenGL ES pipelines, profile and optimize performance on constrained hardware, integrate graphics into the robot platform, collaborate with design/HMI, and establish graphics development best practices and CI.

Own and optimize on-board inference architecture for humanoid robots: map and partition models across heterogeneous accelerators, enforce compute budgets, optimize toolchains and kernels, apply quantization/pruning, profile pipelines, and collaborate with AI, platform, and silicon vendors to meet real-time power, latency, and reliability constraints.

Own end-to-end demand planning, build S&OP and forecasting processes, aggregate commercial and internal requirements, model demand including yield and test consumption, prioritize parts under constraints, create planning metrics and dashboards, and evaluate planning/ERP tools while collaborating with Manufacturing, Quality, NPI, Finance, and Commercial teams.

Develop and maintain XR products and spatial UIs to interface with humanoid robots and data-capture hardware. Implement features, fix bugs, harden XR data sync, integrate custom hardware with XR headsets, and support hardware experiments, low-latency video streaming, and motion-capture workflows.

Lead onsite NPI and board bring-up at JDM partner facilities across Asia. Develop C/C++ diagnostic firmware, integrate and maintain automated factory test fixtures, and debug system-level manufacturing defects using scopes, logic analyzers, and custom tooling. Act as the technical liaison between US R&D and local manufacturing teams to transition prototypes through EVT/DVT/PVT to mass production.

The Helix Data Creator will wear sensors to guide and collect motion data for training humanoid robots, providing feedback and maintaining equipment during data collection.

Optimize distributed training for 100B+ models across massive GPU fleets by writing and tuning custom kernels, improving IO/checkpointing/fault tolerance, building monitoring tooling, co-designing model and training recipes, and benchmarking heterogeneous accelerators and parallelism strategies.