What's Open and What It Signals
A six-person team fresh out of Y Combinator's summer 2026 batch is advertising a $400,000 ceiling for a researcher. Aktoria Robotics lists two roles: Robot Learning Researcher (VLA / Foundation Models) in San Francisco at $200,000–$400,000, Y Combinator's job postings show, and Robotics Operations Lead in Delhi at ₹3.5 million–₹6 million (roughly $42,000–$72,000) with 1% equity. The spread maps the company's bet: autonomy arrives in layers, and the layer paying the bills right now is a global fleet of trained humans driving robots through a 40-millisecond teleoperation link.
Aktoria describes itself as the "teleoperation layer that speeds up robot deployments" — low-latency software paired with a "Robot Operations Center" staffed by in-house operators across the United States, Mexico, and India. Every "autonomous" robot deployment this year had a human behind it, the founders argue, and building an in-house teleop team is the bottleneck between a working prototype and a paying customer. The Delhi role owns that bottleneck. The listing describes a "real robotics operations role: you wear a VR headset, and your movements drive the robot through real environments." The operator captures stereo and IMU egocentric data, handheld UMI recordings, and full teleoperation episodes — datasets that feed back into the foundation models the San Francisco researcher is hired to advance.
The San Francisco role signals where the puck is moving. Vision-language-action models (VLAs) are the current frontier in robot learning. Aktoria's researcher will work on "VLA / Foundation Models" with direct access to the proprietary teleoperation dataset the operations center generates daily. The salary band reaches into staff-engineer territory at top AI labs.
The equity structure tells its own story. The Delhi operations lead receives 1% — meaningful for an early employee in a seed-stage company. The researcher role lists no equity figure in the public postings. The asymmetry mirrors the market: operations talent in India is abundant relative to Bay Area foundation-model researchers, and the company prices accordingly.
What's missing is equally revealing. No mechanical engineering role. No hardware bring-up position. No perception stack engineer. Aktoria is not building robots; it is building the layer that makes other people's robots deployable today while generating the data that may make that layer obsolete tomorrow. The two open roles are the two sides of that wager.
Inside the Screen
Aktoria Robotics does not publish its screening playbook. The company's careers page, its Y Combinator profile, and public founder interviews describe the product, a 24-hour global teleoperation network spanning India, Mexico, and the United States, backed by sub-40-millisecond cross-country latency, but they are silent on the mechanics that separate a resume from a first interview.
The two active listings, Robot Learning Researcher and Robotics Operations Lead, sit in a category that, as of late 2024, represented roughly one-third of hiring at the top robotics companies tracked in public compilations. Research from Aced (formerly Exponent) notes that forward-deployed and deployment engineers now appear at a good one-third of hot AI companies, up from just Palantir in 2018.
Candidates report noping out of processes more frequently when the interview content diverges from the posted requirements — a dynamic the research flags as growing across AI and robotics hiring. For Aktoria, the screen filters for the exact profile the product demands: engineers who have already lived the latency budget, the field deployment, and the 24/7 operator handoff.
Clearing the Bar
The two open roles read like a split-screen view of what Aktoria values. One seat demands research depth at the frontier of vision-language-action models. The other demands operational grit: standing up a 24/7 robot operations center in India from zero, importing hardware, wiring power and network, hiring supervisors and shift operators, and owning the site P&L.
On the research side, the bar is set by the founders themselves. Shreenabh Agrawal, ex-Zoox, carries first-author papers at top robotics conferences. Jaskaran Singh Walia has 15-plus publications spanning MIT, Cambridge, and Microsoft Research. The Robot Learning Researcher role explicitly targets VLA and foundation model expertise. The Y Combinator job posting for the role sits alongside the Operations Lead listing, signaling the company treats both as equally critical to the product — a single loop.
The Operations Lead screen is strict in a different dimension. The job description is unambiguous: first hire in India, build the office, procure and import robots, run fit-out, hire the entire floor across multiple shifts, then own throughput, quality, uptime, retention, and P&L. That is a general-manager mandate wrapped in a robotics-operations title. The founders are explicit that the hire "reports to the founders directly" and "hires everyone who comes after."
Across both roles, the screen rewards comfort with the "human-in-the-loop" reality that Aktoria sells. The company's pitch, that every "autonomous" robot relies on a human operator, is the operating assumption. The founders built the company on that premise; the screen selects for people who want to build on it.
With a six-person team and a YC partner in Diana Hu, the interview loop is short and high-bandwidth. A candidate's ability to communicate technical trade-offs directly to the founders, without layers of management translation, is itself a screened trait.
Boston Dynamics lists a Teleoperations Research Engineer on Atlas at $150k–$200k, confirming the market price for this niche. Apptronik lists a Staff Software Engineer for humanoid controls at $300k–$350k, showing where the ceiling sits.
| Company | Role | Salary Band |
|---|---|---|
| Boston Dynamics | Teleoperations Research Engineer (Atlas) | $150k–$200k |
| Apptronik | Staff Software Engineer – Humanoid Controls | $300k–$350k |
Geography complicates both searches. Aktoria's operator network spans Haryana, India, and Mexico alongside San Francisco. The company's own board data shows one employee in California and one in Haryana; that ratio will shift as the operator headcount grows.
Skill adaptation is visible in public forums. The elevationprovinggrounds.com guide for 2026 hiring flags Digital Twins and Edge AI deployment as the primary differentiators for top-tier robotics talent. Candidates targeting Aktoria are adding Isaac Sim, MuJoCo, and real-to-sim pipelines to portfolios because teleop data collection (cloth folding, cable insertion, resin cleanup) lives or dies on simulation fidelity. LinkedIn's own career advice threads point engineers toward ROS 2, paperswithcode, and Hugging Face model hubs as the self-study path into VLA work. The Udacity Robotics Nanodegree and RoboCup participation show up repeatedly as credible signals for early-career applicants who lack lab publications.
No public write-ups yet describe Aktoria's interview loop in detail. The company is too new, the roles too fresh. Candidates are operating on founder reputation, YC stamp, and the teleop thesis — that every "autonomous" demo depends on a human operator.
The Candidate Response
The two roles sit at opposite ends of the stack. The researcher role targets a skill set that barely existed two years ago: vision-language-action models that turn internet-scale pretraining into robot policies. Aktoria's founders, Walia (15+ papers, MIT/Cambridge/Microsoft Research) and Agrawal (ex-Zoox, first-author at top conferences), are themselves that profile.
The Operations Lead role draws a different pool. Teleoperation at scale, with 24/7 coverage across India, Mexico, and the U.S., and ~30 ms cross-country latency per aktoria.com, requires someone who has built or run distributed human-in-the-loop systems.
Market Signals
The North American robotics market posted a healthy rebound in the first half of 2025 after a slow 2024, but the overall growth figure masks a more consequential shift: the talent pipeline is not keeping pace. Deloitte's 2024 manufacturing study projects 3.8 million net new employees needed by 2033. A Robotics and Automation News report cites over 600,000 unfilled manufacturing jobs converging with a wave of Baby Boomer retirements. By 2031, 41 percent of construction workers are expected to retire while only 10 percent of current workers are under 25. The skilled-craft shortage could exceed two million by 2028 if current trends persist.
At the same time, capital is flooding into the physical layer. Companies have announced more than $500 billion in private commitments to revitalize the US chipmaking ecosystem, targeting a tripling of domestic capacity by 2032 and over 500,000 new jobs. Data-center buildout alone could drive power demand from 33 gigawatts in 2024 to 176 gigawatts by 2035, with AI-specific capacity growing thirtyfold to 123 gigawatts. Construction wages rose 4.2 percent year-over-year as of August 2025, and the engineering and construction sector needs 499,000 new workers by 2026, up from 439,000 in 2025.
Manufacturers are responding. Eighty percent of executives surveyed by Deloitte plan to invest 20 percent or more of improvement budgets in smart manufacturing. Twenty-two percent expect to use physical AI within two years, more than double the 9 percent using it today. Ninety-two percent say smart manufacturing will be the main competitiveness driver over the next three years. The most reported method for building digital capabilities? Hiring new talent: 68 percent. Over half have formed central teams to research and deploy these initiatives.
The hiring market reflects that urgency. Zero G Talent's board shows Zipline adding 19 roles in the past week with a salary band of $62k–$255k (median $186k). Boston Dynamics posted four roles at $78k–$206k (median $157k). Apptronik listed four at $59k–$325k (median $230k), including a staff-level humanoid controls role at $300k–$350k.
The screening rigor at companies like Aktoria mirrors a broader realization: automating broken processes fails. Gartner predicts 40 percent of agentic projects will collapse by 2027 because organizations automate instead of redesign. The winners, such as Amazon with its millionth robot and DeepFleet AI cutting warehouse travel 10 percent, and BMW with cars driving themselves through kilometer-long production routes, pair hardware fluency with process redesign. They hire for the intersection.
That intersection is narrowing. ABI Research analyst Michael Larner warns that improving mentoring and knowledge-sharing matters as much as digital investment. The most strategic leaders are funding hands-on apprenticeship programs that capture tribal knowledge on the floor. Others are building university pipelines before vacancies open. The "build, buy, or borrow" workforce framework is replacing reactive posting.
Aktoria's two-role screen, filtering for sim-to-real transfer, hardware-in-the-loop fluency, and the ability to ship in a regulated environment, is a leading indicator. The market is no longer hiring roboticists who can code. It is hiring engineers who can make autonomy reliable in the physical world. The companies that codify that bar now will own the talent pool when the next capital wave hits.
Working in robotics? Zero G Talent tracks the openings: see every open Zipline role, browse robotics jobs, openings at Boston Dynamics and Apptronik, and the people building the field.