Humanoid robots still rely on people to generate physical demonstrations and supervise unfamiliar situations. These teleoperators form a growing hourly workforce that physical AI companies must plan for as deployments move from controlled tests into commercial use.
Key Takeaways
- A robot teleoperator uses control equipment, such as a teleoperation rig or sensor suit, to guide a robot or generate physical demonstrations for its training system.
- Figure's current postings advertise $30 to $32 an hour for data-collection and robot-operation roles, with its Helix Data Creator night shift listed at $33 an hour.
- In Figure's current hiring model, data creators produce training demonstrations, while robot operators supervise commercial deployments and report failures to engineering.
- Companies that have not defined either role can struggle to estimate the staffing required for deployment.
Teleoperation and robot operation serve related functions across training and deployment. Teleoperators usually produce demonstrations, while robot operators often supervise commercial sessions and report failures to engineering. Together, these positions form an hourly workforce that companies must account for when scaling physical AI programs.
Robot Teleoperator vs. Robot Operator
Companies do not use "robot teleoperator" and "robot operator" consistently. In Figure's current hiring model, however, the positions cover different stages of robot training and deployment.
[Table 1]
Figure's Humanoid Robot Pilots and Helix Data Creators focus on producing training demonstrations. They put on a control rig or sensor suit and physically perform a task, over and over, so the robot has real movement to learn from. Figure's Humanoid Robot Pilot listing puts this role at $32 an hour. Its Helix Data Creator posting runs a 24/7 rotation across day, swing, and night shifts, starting at $30 an hour with a $3 differential overnight.
Figure's Humanoid Robot Operator posting, also at $32 an hour, describes staff who deploy robots for commercial use, monitor sessions, and file issues to engineering in real time. Figure treats this as a separate operational function from the roles focused primarily on training data. The titles aren't standardized across the industry, and teleoperation can also support live intervention rather than data collection.
Figure advertises its Humanoid Robot Pilot and Helix Data Creator positions as fixed-term hourly roles. Its Humanoid Robot Operator posting describes a contract-to-hire arrangement, and several positions follow shift-based schedules.
These jobs can also serve as an entry point into robotics. Figure describes its Humanoid Robot Pilot position as a six-month fixed-term role and says high-performing employees can advance into Senior Pilot or other lead positions. Previous robotics experience appears under bonus qualifications, while the core requirements emphasize physical coordination, spatial awareness, sustained focus, and the ability to learn new physical tasks. That profile opens physical AI work to candidates outside the traditional engineering pipeline.
How Robot Teleoperation Works
A teleoperation session follows a repeatable sequence, whether the setup uses VR headsets, hand-tracking rigs, or full-body sensor suits. The operator receives feedback from the robot's sensors and uses control equipment to guide its movements. Depending on the system, the interface can include video feeds, tracked controllers, a headset, or a full-body sensor suit.
The robot mirrors the operator's motion while the system records the session. NVIDIA's documentation for its whole-body control platform describes the robot tracking the operator's movements in real time, and it calls for a second person, a safety operator, stationed at the controls with the ability to trigger an immediate stop if something goes wrong. Fourier's open-source teleoperation system describes a similar setup: one person, the pilot, wears the headset and controls the robot, while a second person, the operator, monitors the robot's position and manages the session.
Both documented systems assign a second person to monitor safety or manage the session while the pilot controls the robot. This setup shows why staffing calculations must account for everyone supporting a teleoperation session.
During collection, the operator can keep or discard an episode. Fourier's system allows an unsuccessful attempt to be removed before it moves further through the data pipeline.
Teleoperation also serves a second purpose beyond data collection. The same skill set applies when a deployed robot's autonomy fails to complete a task and a person needs to step in and finish it manually, which is part of why the line between a teleoperator training a model and a robot operator keeping one running isn't always sharp in practice.
Why This Workforce Exists
Manipulation policies can learn from human demonstrations. For physical tasks, those demonstrations provide examples of the motions required under realistic operating conditions. Teleoperators repeat the movements so the system can record task-specific training data.
That data can bring a model to a working baseline. Commercial sessions then introduce materials and operating conditions the system may not have encountered during training. Staff prepare the work area, monitor the robot, intervene when necessary, and document unfamiliar cases. Required headcount depends on session volume, task complexity, and the system’s current autonomy level.
During early deployment, staffing demand can grow as companies add workflows and encounter new materials or failure conditions. A plant using robots across several workflows may need task-specific demonstrations and supervised testing until each application reaches the required performance level.
What Makes a Good Robot Teleoperator
Figure's listings emphasize operational and physical abilities over formal technical credentials. The postings ask for physical coordination and spatial awareness, the ability to sustain concentration across a repetitive eight-hour shift, and the discipline to execute the same movement the same way every time.
Consistency is central to the role. Figure’s Helix Data Creator listing requires employees to follow “Ground Truth” movement protocols without deviation, work toward daily collection targets, and repeat physical movements for shifts lasting eight hours or longer.
Safe equipment handling also matters. NVIDIA’s documentation requires a clear safety zone and a second operator prepared to trigger an emergency stop. Figure asks its Pilots to maintain a safe environment and follow robot startup, shutdown, and maintenance procedures.
Figure does not require previous robotics experience for its Humanoid Robot Pilot position. Robotics experience and mechanical troubleshooting appear as bonus qualifications, while the core criteria focus on physical coordination, sustained attention, and accurate execution. Gaming or VR experience shows up in the same bonus category on Figure's Helix listing, which is a reasonable proxy: someone with precise controller input or spatial tracking experience from gaming is already partway to the hand-eye coordination the job asks for.
That’s the profile Figure is hiring for right now, and it reaches candidates outside the traditional robotics engineering pipeline.
What This Means for Hiring
Companies scaling a physical AI deployment need written qualification standards for these positions. Relevant skills include reading physical workflows accurately, following operating procedures, identifying abnormal robot behavior, and producing consistent demonstrations that create usable data. Without a defined position, companies can end up assigning the work to engineers already responsible for developing the system.
Before opening these positions, employers should decide whether each role centers on training-data collection, commercial supervision, or a combination of both. That choice affects scheduling, physical requirements, training time, and escalation authority. Data Creators need consistency across repeated movements, while commercial operators need the judgment to recognize failures and report them clearly. Defining the function first gives hiring teams a more accurate candidate profile.
I wrote about the five leadership roles physical AI production creates, including a workforce transition leader responsible for this type of staffing plan. Teleoperators and robot operators are part of the operating workforce that leader must anticipate. Christian & Timbers recruits physical AI executives who can build the organization required to move robotics programs into production. Our executive search team works across physical AI, robotics, autonomous systems, and advanced manufacturing.
Frequently Asked Questions
- What does a robot teleoperator do?
A robot teleoperator uses control equipment, such as a teleoperation rig or sensor suit, to guide a robot or generate the physical demonstrations that train its manipulation policy.
- Is a robot teleoperator the same as a robot operator?
The titles sometimes overlap. In Figure's current postings, Pilots and Helix Data Creators generate training demonstrations, while Humanoid Robot Operators monitor robots in commercial use and report failures to engineering.
- How much does a robot teleoperator make?
Figure's current postings advertise $30 to $32 an hour for data-collection and robot-operation roles, with its Helix Data Creator night shift listed at $33 an hour.
- Do robot teleoperators need engineering or robotics experience?
Figure does not list previous robotics experience as a core requirement for its Humanoid Robot Pilot role. The company emphasizes physical coordination, spatial awareness, sustained concentration, and the ability to learn new movements quickly. Experience with robots or mechanical troubleshooting is listed as an additional qualification.
- Who should plan staffing for robot teleoperators?
A workforce transition leader can coordinate the staffing model with site operations. The appropriate owner depends on the company’s organizational structure, but responsibility should be assigned before deployment expands.