Confidential client

Research Scientist, Robot LearningSan Francisco, CA

Posted

Work style
Onsite
Salary
$500k–$800k
Stage
Seed
Industry
Robotics
Visa sponsorship
Limited
Employment
Full-time

The opportunity

Set the research agenda at a venture-backed robotics research and data company: own questions from hypothesis to result across simulation and real robots, and turn findings into training data and systems at scale. Onsite in San Francisco.

A venture-backed robotics research and data company working with leading robotics labs on how machines learn from the physical world. Its work spans real-world data collection, simulation, training and evaluation, with a focus on generalization, learning from fewer demonstrations, in-context learning and recovery. It runs a very large network of data contributors and works directly with the founders on research direction.

What you will do

  • Work with partner labs to find where robot models fall short, and pick the research questions that matter most
  • Design controlled experiments and ablations across simulation and real-world systems
  • Train and debug robot policies, proxy models, and models that predict transfer or the value of data
  • Build reproducible pipelines for dataset generation, training, simulation, evaluation and analysis
  • Show why an approach works, where it fails, and whether the gains hold across models and settings
  • Design collection methods across task and environment mixes, camera views, sensors and demonstration sources
  • Turn validated findings into scalable collection and training processes with engineering and operations
  • Explore directions such as data value and transfer, sim-to-real, weight-space edits to VLA models, failure-driven recovery, and learning from human video, teleoperation and UMI

What you bring

  • Hands-on experience training and debugging robot policies: VLAs, diffusion policies, imitation learning or reinforcement learning
  • Strong experimental judgment: original hypotheses, careful handling of confounders, tests that settle the question
  • Experience with manipulation, simulation, real-world evaluation or related embodied learning systems
  • Strong programming and the ability to build reliable training and research infrastructure
  • A solid grasp of generalization, distribution shift and transfer between settings
  • Evidence of independent technical thinking and ownership through research, projects or experiments

Nice to have

  • Sim-to-real transfer, world models, meta-learning, data selection or synthetic data
  • Policy adaptation or large-scale training
  • Robotics simulators, manipulation hardware or multimodal datasets
  • Familiarity with UMI, ALOHA, DROID, Open X-Embodiment or learning from human video

Pay and benefits

$500k–$800ka year, full-time

  • Salary range: $500,000 to $800,000 per year. This is the range the client expects to pay in good faith. The final amount depends on experience and other job-related factors.
  • Total compensation of $500K to $800K depending on experience and scope
  • Equity and direct work with the founders on research direction

Where you will work

The team is based in San Francisco, CA, and works onsite. We can talk about what that looks like for you.

This role has limited visa sponsorship. We will confirm what is possible for your situation.

Why is the company hidden? Our clients hire quietly. V6 shares the name once we have spoken and you are ready.

How we hire

  1. ApplyTwo minutes. A resume or pasted text, plus a few lines on why you fit.
  2. V6 screenA practitioner reviews your application, and we call when there's a match.
  3. Intro call
  4. Technical conversations with the founders
  5. References
  6. Offer

Interview steps are set by the client and may change. V6 does not control the client's process, but we advocate for every candidate at each stage.

About this listing

V6 posts this role on behalf of a client company, which makes the hiring decision. Role details, salary range and benefits come from the client and may change or close without notice. A listing is not an offer of employment, and applying does not create an employment, agency or contractual relationship with V6 or the client. Candidates never pay V6 any fee.

Equal opportunity. V6 and our clients consider all qualified applicants without regard to race, color, religion, sex, gender identity or expression, sexual orientation, national origin, ancestry, age, disability, medical condition, genetic information, marital status, military or veteran status, or any other characteristic protected by law. Qualified applicants with arrest or conviction records will be considered in line with the California Fair Chance Act. Need an accommodation to apply? Email [email protected].

Research Scientist, Robot LearningApply