EmergingComputer vision and spatial perception

Open-World Perception Researcher

Also called Open-Vocabulary Vision Scientist

An Open-World Perception Researcher builds vision models that know when they are looking at something unfamiliar. That matters because a system that guesses confidently on unknown objects creates risk for your product and your users.

What they do

They take a perception model trained on a fixed set of labels and test what happens when the world goes beyond it. They build evaluation sets of unknown objects, design experiments to separate real novelty from ordinary noise, and compare methods for open-set recognition and uncertainty estimation. They train and adjust deep learning models in PyTorch, use OpenCV for data handling, and study failure cases by hand. They also decide with product teams when the model should abstain, flag, or hand over to a person, and write up the trade-offs clearly.

What a strong hire delivers

In the first months, a strong hire gives you a perception model with unknown-class testing built in, not added afterwards. They set up an evaluation that measures unknown-object detection and the quality of abstention. They show where the current system is overconfident, fix the worst cases, and leave the team with a repeatable test process and a clear view of residual risk.

How you can tell they are good

Ask them to walk through a model they made safer on unfamiliar inputs. Listen for how they built the unknown test set, which metrics they chose and why, and what they would not trust. Strong people talk about calibration, thresholds and trade-offs between missed unknowns and needless abstention. Be wary of anyone who only quotes accuracy on known classes.

Why it is hard to fill

Few people carry this title. The work often sits inside computer vision, robotics or safety roles, so CVs rarely say open-set or open-world. Many candidates know detection well but have never designed unknown-class evaluation. Others know uncertainty theory but have little hands-on vision experience. You are looking for the overlap, and it is thin. Expect to judge on the work shown, not the job title.

Where to find them

Look at computer vision researchers, applied ML engineers and image processing specialists who have worked on anomaly detection, out-of-distribution detection or uncertainty estimation. They sit in robotics, autonomous systems, inspection and imaging teams, and in research groups at universities and industrial labs. Some are in safety or reliability roles where abstention matters. Publications and open code on open-set recognition are a useful signal.

Indicative pay

Not enough published evidence yet to give a fair range for this role, which is common for work this new. We benchmark it for you as part of a search.

Interview questions for an Open-World Perception Researcher

  1. Describe a vision model you made behave better on objects it had never seen. What did you change and how did you know it worked?

    A strong answer names the unknown test set, the metrics and the before and after result. It covers what was tried and dropped. Worry if the answer rests on known-class accuracy or on vague claims of improved robustness.

  2. How would you build an evaluation set of unknown objects when you cannot list everything the model might meet?

    Good answers hold out whole classes, vary context and difficulty, and separate novel objects from image corruption. They admit no set is complete. Worry if they treat one held-out split as proof of safety.

4 more questions, with what a strong answer covers, are in the Role Kit.

Job description you can post

We are hiring an Open-World Perception Researcher to make our vision systems recognise the limits of what they know. You will build and test perception models that detect unfamiliar objects and abstain safely, working closely with engineering and product to turn research into reliable behaviour.

What you will do

  • Develop and train perception models that detect unknown objects outside closed label sets
  • Design evaluation sets and experiments for unknown-class testing
  • Measure and improve unknown-object detection and abstention quality
  • Apply uncertainty quantification and calibrate model confidence
  • Analyse failure cases and report findings clearly to the team
  • Agree abstention and handover behaviour with product colleagues
  • Document methods so tests can be repeated

What you will bring

  • Demonstrated work in computer vision, image processing or applied ML
  • Hands-on experience with open-set recognition, out-of-distribution detection or similar
  • Strong deep learning skills in PyTorch
  • Experience with OpenCV or comparable image tooling
  • Sound experimental design, with examples of evaluations you built
  • Practical grasp of uncertainty quantification and calibration
  • Clear written communication of results and limits

Nice to have

  • Published or open-source work on open-world or open-vocabulary vision
  • Experience in robotics, inspection or other safety-sensitive systems
  • Experience with vision-language models in open-vocabulary settings

Free to use and adapt. If it saves you time, a link back to this page is appreciated.

Skills that matter

  • Computer visionAdvancedAnalyse or generate visual information
  • Deep learningAdvancedBuild and diagnose trainable neural models
  • Experimental designWorkingChoose comparisons that identify credible effects
  • Uncertainty quantificationWorkingEstimate confidence, calibration and model uncertainty

Other roles in computer vision and spatial perception

Hiring nowComputer Vision EngineerA Computer Vision Engineer builds detection and recognition systems that work in a defined setting, and proves how well they work. Companies rely on…Hiring now3D Vision ResearcherA 3D Vision Researcher recovers spatial structure from images and depth data, then proves it holds up. Companies building robotics, mapping, AR or…Hiring nowVideo Understanding ScientistA Video Understanding Scientist builds models that recognise actions, events and the dependencies between them in video. Companies hire for this…

Related roles in other families

Hiring nowScientific Machine Learning ScientistScientific machine learning and research computingA Scientific Machine Learning Scientist builds learned models that obey the physics, chemistry or biology of a problem. They matter because a model…Hiring nowAI Research Software EngineerScientific machine learning and research computingAn AI Research Software Engineer turns scientific machine learning ideas into software that other people can run, trust and extend. They give a…EmergingAI Scientific Workflow EngineerScientific machine learning and research computingAn AI Scientific Workflow Engineer connects data, models and instruments into reproducible science pipelines. They make sure results can be traced…EmergingAI Reproducibility SpecialistScientific machine learning and research computingAn AI Reproducibility Specialist independently reruns AI-supported scientific claims and reports what holds and what does not. They protect the…

How we wrote this

Role definitions are our own editorial work, checked against occupational classifications, standards, research and employer pages. Titles vary by company, and this is a description of the work, not a job advert.

← Computer vision and spatial perceptionAll new job families

Start a search

Hiring an Open-World Perception Researcher?

Tell us the work. We find the people already doing it, whatever their job title says, and you meet them already interviewed for your role.