Key findings
- 2
122 open roles state no arrangement in their title, location or workplace field, 35.2% of them at Avride, so the remote share is a floor.
- 3
AV & Drone Operators has the largest on-site share of any kind of work with at least 10 open roles (86.9%, 305 roles); 7 kinds of work have any on-site roles at all.
- 4
35 employers list on-site roles; TSMG lists the most (474, 81.6% of all on-site roles).
- 5
On-site roles are open in 49 countries; the United States has the most (230, 39.6%), and the place named most is the San Francisco Bay Area (40 roles).
- 6
Read from their titles, the largest groups of on-site roles are autonomous vehicle and truck operation (297 roles, 51.1%) and robot data collection (147, 25.3%).
Figure 1
On-site, remote or not stated, by kind of work
Which kinds of robot training work have to be done at a site rather than from home?
| Kind of work | On-site or hybrid | Remote | Not stated |
|---|---|---|---|
| All kinds of work | 78.0% | 5.6% | 16.4% |
| AV & Drone Operators | 86.9% | 0.3% | 12.8% |
| Data & Demonstration Collection | 85.7% | 6.3% | 7.9% |
| Robot Teleoperation | 80.2% | 0.0% | 19.8% |
| Egocentric & Wearable Capture | 77.6% | 22.4% | 0.0% |
| Robot Fleet Operations | 52.5% | 0.0% | 47.5% |
| Embodied-AI Data Annotation | 20.0% | 40.0% | 40.0% |
| Robot Task Evaluation & QA | 18.2% | 45.5% | 36.4% |
Sample. 745 roles open at the end of 9 Oct 2026; 581 on-site or hybrid.
On-site or hybrid: the title or location names a site ('Onsite', 'On-site London', 'In-person', 'Hybrid'), or a job board that publishes a workplace field (Lever, Ashby) marks the role not remote. Remote: the title or location says remote, or the board flags the role remote. 'Remote' that names the work ('Remote Driver', 'Remote Assistance Operator') says how a robot or vehicle is controlled, not where the worker sits, and is not read as remote.
Not stated: neither the title, the location nor a workplace field says how the work is done. Most robot training roles are tied to a lab, a depot or a test route, but a missing word is not read as on-site, so both the on-site and the remote shares are floors, not estimates of the rest.
Hybrid roles are counted with on-site ones: the listing fields do not separate them.
Source. RobotTraining.jobs Research, On-Site Robot Training Work, live data.
Figure 2
On-site, remote or not stated, by employer
Which employers list robot training work that has to be done in person?
| Employer | On-site or hybrid | Remote | Not stated |
|---|---|---|---|
| TSMG | 100.0% | 0.0% | 0.0% |
| Serve Robotics | 100.0% | 0.0% | 0.0% |
| Aurora Innovation | 100.0% | 0.0% | 0.0% |
| Dyna RoboticsSmall sample | 100.0% | 0.0% | 0.0% |
| WaabiSmall sample | 100.0% | 0.0% | 0.0% |
| ISEESmall sample | 100.0% | 0.0% | 0.0% |
| SaronicSmall sample | 100.0% | 0.0% | 0.0% |
| Collaborative RoboticsSmall sample | 100.0% | 0.0% | 0.0% |
| Genesis AISmall sample | 100.0% | 0.0% | 0.0% |
| WayveSmall sample | 100.0% | 0.0% | 0.0% |
| DexmateSmall sample | 100.0% | 0.0% | 0.0% |
| FS StudioSmall sample | 100.0% | 0.0% | 0.0% |
| SereactSmall sample | 100.0% | 0.0% | 0.0% |
| Shield AISmall sample | 100.0% | 0.0% | 0.0% |
| micro1 | 10.0% | 80.0% | 10.0% |
| PykaSmall sample | 66.7% | 33.3% | 0.0% |
| Sunday RoboticsSmall sample | 66.7% | 33.3% | 0.0% |
| 1X TechnologiesSmall sample | 100.0% | 0.0% | 0.0% |
| Cota RoboticsSmall sample | 100.0% | 0.0% | 0.0% |
| HumanoidSmall sample | 100.0% | 0.0% | 0.0% |
| May MobilitySmall sample | 100.0% | 0.0% | 0.0% |
| MonumentalSmall sample | 100.0% | 0.0% | 0.0% |
| Rhoda AISmall sample | 100.0% | 0.0% | 0.0% |
| Tutor IntelligenceSmall sample | 100.0% | 0.0% | 0.0% |
| NuroSmall sample | 16.7% | 0.0% | 83.3% |
| AdexisSmall sample | 100.0% | 0.0% | 0.0% |
| Agility RoboticsSmall sample | 100.0% | 0.0% | 0.0% |
| AppenSmall sample | 100.0% | 0.0% | 0.0% |
| Generalist AISmall sample | 100.0% | 0.0% | 0.0% |
| Havoc AISmall sample | 100.0% | 0.0% | 0.0% |
| Mind RoboticsSmall sample | 100.0% | 0.0% | 0.0% |
| Physical IntelligenceSmall sample | 100.0% | 0.0% | 0.0% |
| Reliable RoboticsSmall sample | 100.0% | 0.0% | 0.0% |
| SkydioSmall sample | 100.0% | 0.0% | 0.0% |
| SkywaysSmall sample | 100.0% | 0.0% | 0.0% |
| Avride | 0.0% | 0.0% | 100.0% |
| Claru | 0.0% | 100.0% | 0.0% |
| Gatik | 0.0% | 0.0% | 100.0% |
| WaymoSmall sample | 0.0% | 55.6% | 44.4% |
| Figure AISmall sample | 0.0% | 0.0% | 100.0% |
| ZiplineSmall sample | 0.0% | 0.0% | 100.0% |
| Anduril IndustriesSmall sample | 0.0% | 0.0% | 100.0% |
| Kodiak AISmall sample | 0.0% | 0.0% | 100.0% |
| Skild AISmall sample | 0.0% | 0.0% | 100.0% |
| Nimble RoboticsSmall sample | 0.0% | 0.0% | 100.0% |
| PerleSmall sample | 0.0% | 100.0% | 0.0% |
| Stack AVSmall sample | 0.0% | 0.0% | 100.0% |
| Bot AutoSmall sample | 0.0% | 0.0% | 100.0% |
| Keywords StudiosSmall sample | 0.0% | 0.0% | 100.0% |
| NerosSmall sample | 0.0% | 0.0% | 100.0% |
| RoboForceSmall sample | 0.0% | 0.0% | 100.0% |
| VaySmall sample | 0.0% | 0.0% | 100.0% |
| Allen Control SystemsSmall sample | 0.0% | 0.0% | 100.0% |
| DexorySmall sample | 0.0% | 0.0% | 100.0% |
| Maven RoboticsSmall sample | 0.0% | 0.0% | 100.0% |
| MotionalSmall sample | 0.0% | 0.0% | 100.0% |
| ProntoSmall sample | 0.0% | 0.0% | 100.0% |
| SamaSmall sample | 0.0% | 0.0% | 100.0% |
| Torc RoboticsSmall sample | 0.0% | 0.0% | 100.0% |
| WingSmall sample | 0.0% | 0.0% | 100.0% |
Sample. 745 roles open at the end of 9 Oct 2026, by the employer that lists them (listings, not hires).
On-site or hybrid: the title or location names a site ('Onsite', 'On-site London', 'In-person', 'Hybrid'), or a job board that publishes a workplace field (Lever, Ashby) marks the role not remote. Remote: the title or location says remote, or the board flags the role remote. 'Remote' that names the work ('Remote Driver', 'Remote Assistance Operator') says how a robot or vehicle is controlled, not where the worker sits, and is not read as remote.
Not stated: neither the title, the location nor a workplace field says how the work is done. Most robot training roles are tied to a lab, a depot or a test route, but a missing word is not read as on-site, so both the on-site and the remote shares are floors, not estimates of the rest.
Source. RobotTraining.jobs Research, On-Site Robot Training Work, live data.
Figure 3
Where on-site robot training roles are, by country
In which countries is in-person robot training work open?
| Country | On-site roles | Share of on-site roles |
|---|---|---|
| United States | 230 | 39.6% |
| Japan | 33 | 5.7% |
| Germany | 30 | 5.2% |
| Canada | 26 | 4.5% |
| Switzerland | 23 | 4.0% |
| Spain | 20 | 3.4% |
| Italy | 19 | 3.3% |
| United Kingdom | 16 | 2.8% |
| Philippines | 14 | 2.4% |
| France | 13 | 2.2% |
| Netherlands | 13 | 2.2% |
| Indonesia | 12 | 2.1% |
| Malaysia | 12 | 2.1% |
| Australia | 10 | 1.7% |
| Belgium | 10 | 1.7% |
| Thailand | 8 | 1.4% |
| Sweden | 7 | 1.2% |
| Vietnam | 7 | 1.2% |
| Norway | 6 | 1.0% |
| Austria | 5 | 0.9% |
| Cambodia | 5 | 0.9% |
| Poland | 5 | 0.9% |
| Finland | 4 | 0.7% |
| Bulgaria | 3 | 0.5% |
| Denmark | 3 | 0.5% |
| Estonia | 3 | 0.5% |
| Luxembourg | 3 | 0.5% |
| Portugal | 3 | 0.5% |
| Romania | 3 | 0.5% |
| Chile | 2 | 0.3% |
| Cyprus | 2 | 0.3% |
| Greece | 2 | 0.3% |
| Croatia | 2 | 0.3% |
| Hungary | 2 | 0.3% |
| Ireland | 2 | 0.3% |
| Lithuania | 2 | 0.3% |
| Latvia | 2 | 0.3% |
| New Zealand | 2 | 0.3% |
| Saudi Arabia | 2 | 0.3% |
| Singapore | 2 | 0.3% |
| Slovenia | 2 | 0.3% |
| Slovakia | 2 | 0.3% |
| United Arab Emirates | 1 | 0.2% |
| Colombia | 1 | 0.2% |
| Czechia | 1 | 0.2% |
| Ecuador | 1 | 0.2% |
| Peru | 1 | 0.2% |
| Ukraine | 1 | 0.2% |
| South Africa | 1 | 0.2% |
| Not stated | 3 | 0.5% |
Sample. 581 on-site or hybrid roles open at the end of 9 Oct 2026, by the country their location names.
Read from the location text the posting publishes. A role open at sites in two countries counts once in each, so shares can add up to more than 100%.
A city's country is taken from the location itself (the country or US state beside it, or the one country its list names). Two capital letters after a city are read as a US state. A city that cannot be matched is listed with no country.
Source. RobotTraining.jobs Research, On-Site Robot Training Work, live data.
Figure 4
Cities with on-site robot training roles
Which cities have the most in-person robot training work?
| City or metro area | Country | On-site roles | Naming only this city |
|---|---|---|---|
| San Francisco Bay Area | United States | 40 | 38 |
| Dallas | United States | 16 | 16 |
| London | United Kingdom | 12 | 11 |
| Austin | United States | 11 | 11 |
| Atlanta | United States | 10 | 9 |
| Los Angeles | United States | 9 | 9 |
| Munich | Germany | 9 | 9 |
| Paris | France | 9 | 9 |
| Tokyo | Japan | 8 | 5 |
| Amsterdam | Netherlands | 7 | 7 |
| Barcelona | Spain | 7 | 7 |
| Berlin | Germany | 7 | 7 |
| Pittsburgh | United States | 7 | 7 |
| Stockholm | Sweden | 7 | 7 |
| Brussels | Belgium | 6 | 6 |
| Customer Site | United States | 6 | 6 |
| Houston | United States | 6 | 6 |
| Madrid | Spain | 6 | 6 |
| Miami | United States | 6 | 6 |
| Phoenix | United States | 6 | 6 |
| Stuttgart | Germany | 6 | 6 |
| Bern | Switzerland | 5 | 5 |
| Columbus | United States | 5 | 5 |
| Geneva | Switzerland | 5 | 5 |
| Nashville | United States | 5 | 5 |
Sample. 574 of the 581 on-site roles open at the end of 9 Oct 2026 name a city; the 25 cities named most.
A role that may be worked at any of several sites counts once in each city, so 'Naming only this city' counts the roles tied to that city alone.
Cities of one metropolitan area are merged: Sunnyvale, Burlingame, Santa Clara and San Francisco count as the San Francisco Bay Area; Redmond and Bellevue as the Seattle area; Brooklyn as New York City.
Source. RobotTraining.jobs Research, On-Site Robot Training Work, live data.
Figure 5
What on-site robot training roles are
What do the robot training roles that must be done in person involve?
| Work, read from the title | On-site roles | Share of on-site roles |
|---|---|---|
| Autonomous vehicle and truck operation (safety drivers, test operators) | 297 | 51.1% |
| Robot data collection and demonstrations | 147 | 25.3% |
| Robot service, field and fleet operations | 53 | 9.1% |
| Robot teleoperation and piloting | 47 | 8.1% |
| Other | 14 | 2.4% |
| Drone and flight operations | 12 | 2.1% |
| Robot operation (other) | 9 | 1.5% |
| Robot testing, evaluation and QA | 2 | 0.3% |
Sample. 581 on-site or hybrid roles open at the end of 9 Oct 2026.
Each role counts once, in the first group whose words its title contains, in the order listed in the method (vehicles first, then drones and flight, teleoperation, data collection, service and fleet work, testing, other robot operation).
Source. RobotTraining.jobs Research, On-Site Robot Training Work, live data.
Figure 6
Advertised pay of on-site and remote roles, within each kind of work
Does in-person robot training work advertise more than remote work of the same kind?
| Kind of work | On-site median (per hour) | Remote median (per hour) | On-site against remote |
|---|---|---|---|
| AV & Drone OperatorsSmall sample | $35 | n/a | n/a |
| Robot Fleet OperationsSmall sample | $22 | n/a | n/a |
| Data & Demonstration CollectionSmall sample | $28 | n/a | n/a |
| Robot TeleoperationSmall sample | $26 | n/a | n/a |
| Egocentric & Wearable CaptureSmall sample | n/a | $10 | n/a |
| Embodied-AI Data AnnotationSmall sample | n/a | n/a | n/a |
| Robot Task Evaluation & QASmall sample | n/a | n/a | n/a |
Sample. Roles open at the end of 9 Oct 2026 that state an hourly rate, in the kinds of work with on-site roles.
Medians need at least 8 roles stating a rate on each side; the comparison is made within each kind of work so the mix of work is not read as a premium.
On-site roles advertising an annual salary, a fee per session or no figure are not in the medians.
Source. RobotTraining.jobs Research, On-Site Robot Training Work, live data.
Figure 7
On-site roles over the window, like for like
Is in-person robot training work growing or shrinking?
| Work arrangement | Open on 12 Sep 2026 | Open on 9 Oct 2026 | Change |
|---|---|---|---|
| On-site or hybridSmall sample | 0 | 0 | n/a |
| RemoteSmall sample | 0 | 0 | n/a |
| Not statedSmall sample | 0 | 0 | n/a |
| All rolesSmall sample | 0 | 0 | n/a |
Sample. Roles open at the start of 12 Sep 2026 and at the end of 9 Oct 2026 on the employers tracked throughout; new listings over the whole window.
Open counts compare only employers tracked before the window opened, so an employer added during it does not read as growth.
New listings: roles first seen in the window on an employer already tracked before that week (an employer's first week is its back catalogue).
Source. RobotTraining.jobs Research, On-Site Robot Training Work, live data.
For robotics labs, fleet operators, data vendors and researchers
Put this data to work
A custom cut
Any figure for your countries, cities, kinds of work, vehicles or dates.
A regular feed
The same figures weekly or monthly, as files or through the data endpoints.
Your company, benchmarked
Your listings, pay and reach against the market, from public job boards.
Method and definitions
The public job boards of the robot training employers and programs this site tracks, read every day: each open role's title, its location text and the workplace field its board publishes. Every figure is computed by one database query over the window shown, the same query for the live report and for every edition. Each figure states its sample beneath it.
- On-site or hybrid
- The role must be done at a named place at least part of the week: its title or location says so ('Onsite', 'On-site London', 'In-person', 'Hybrid'), or a job board that publishes a workplace field of its own (Lever, Ashby) marks it not remote. Hybrid roles are counted here because the listing fields do not separate them from on-site ones.
- Remote
- The role's title or location says remote (or work from home, worldwide), or its board flags it remote. 'Remote' that names the work ('Remote Driver', 'Remote Assistance Operator') says how a robot or vehicle is controlled, not where the worker sits, and is not read as remote.
- Not stated
- Neither the title, the location nor a workplace field says how the work is done. A missing word is not read as on-site.
- City or metro area
- Read from the location text: a city's country is the country or US state beside it, or the one country its list names, never guessed. Cities of one metropolitan area are merged. A role open at several sites counts once in each.
- Work, read from the title
- Each on-site role falls in the first of these groups whose words its title contains: autonomous vehicle and truck operation; drone and flight operations; robot teleoperation; robot data collection and demonstrations; robot service, field and fleet operations; robot testing, evaluation and QA; other robot operation; other.
- Advertised pay
- The hourly figure a posting states (the middle of a stated range), compared within each kind of work. Medians need at least 8 roles stating a rate on each side.
- New listing
- A role that first appeared during the window at an employer already tracked before that week. An employer's first week in this dataset is its back catalogue and is not counted as new.
- Open role
- A role listed by a given day and still in its employer's feed that day.
- Window
- Live: the last 28 complete days, in UTC. Edition: one calendar month, in UTC.
What this data cannot tell you
- Both shares are floors: a role that requires on-site sessions may say so only in its description.
- Hybrid and fully on-site roles cannot be told apart from the listing fields.
- Location text says where a role is worked or where its applicants must live, and the two are not always distinguishable; two capital letters after a city are read as a US state.
- Employers that recruit without a public job board are not counted.
- A listing is not a hire, and some employers post one listing per city or site, which weighs a city with many such listings more.
How the job feeds are read and roles are classified: site methodology. How fresh, complete and consistent the data is, checked every day: data quality and coverage.
Editions
The live report covers the last 28 complete days and changes every day. Each complete month is also published as an edition, frozen two days after the month ends, so a figure you cite from an edition never moves.
- Live dataLast 28 complete days, recomputed daily
Cite and reuse
Free to quote, chart and republish under CC BY 4.0 with a link to this page. Live figures change daily: cite a monthly edition for numbers that will read the same later, or keep the access date.
Citation
RobotTraining.jobs (2026). On-Site Robot Training Work (live data, 12 Sep 2026 to 9 Oct 2026). https://robottraining.jobs/research/on-site-robot-training-work (accessed 10 Oct 2026)
BibTeX
@misc{robottrainingjobs_on_site_robot_training_work_live,
author = {{RobotTraining.jobs}},
title = {On-Site Robot Training Work (live data, 12 Sep 2026 to 9 Oct 2026)},
year = {2026},
howpublished = {\url{https://robottraining.jobs/research/on-site-robot-training-work}},
note = {Accessed 10 Oct 2026},
}Data: all figures as CSV (one line per value) or JSON. Every figure also downloads on its own. More open data: datasets and API.
More research
- How Robot Training Roles Advertise PayHow often do robot training roles state their pay, how are the rates distributed and how wide are the ranges?Read the report
- Where Robot Training Work Is OpenWho may apply to robot training roles, country by country?Read the report
- Robot Training Pay TrendsIs the advertised rate for each kind of robot training work rising or falling?Read the report
- What Raises Robot Training PayWhich traits of a robot training role come with higher advertised pay, compared within the same kind of work?Read the report
- How Robot Training Work Is PaidHow is robot training work paid: by the hour, by the task or video, or by salary, on what contract, and how often?Read the report
- The Robot Training Jobs IndexIs the market for human robot training work growing, week by week?Read the report
- The State of Robot Training PayWhat does robot training work pay, by kind of work and by employer?Read the report