Hiring nowAI for physics, mathematics and astronomy

Physics-Informed ML Scientist

Also called Physics ML Researcher

A Physics-Informed ML Scientist builds learned models that respect physical laws, so a company can run fast, trustworthy approximations of costly simulations and trust the results enough to base research or product decisions on them.

What they do

Takes a physical system, such as a fluid flow, a field equation or an astronomical signal, and builds a learned model that stays consistent with the known physics. Chooses how to encode constraints, whether in the loss, the architecture or the data. Trains surrogates against simulation codes and observational data. Designs tests for conservation laws, symmetries and behaviour outside the training range. Quantifies uncertainty. Works with domain scientists to decide what counts as correct, and writes up where the model can and cannot be trusted.

What a strong hire delivers

Within the first months, a strong hire delivers a working surrogate or learned approximation for one real problem, with a clear test suite for physical consistency. They set out accuracy on unseen conditions and measured conservation-law violations. They leave behind a repeatable validation method, documented limits, and a shared view with domain colleagues of when the model is safe to use.

How you can tell they are good

Ask for one model they built and how they knew it was right. Good candidates talk about failure first: where the surrogate broke, which constraint mattered, how they tested outside the training range. They can explain why they chose a soft or hard constraint, and they report uncertainty without being pushed. Be wary of anyone who only quotes in-sample error.

Why it is hard to fill

The title is new and rarely used the same way twice. Candidates tend to be strong physicists who learned machine learning, or machine learning engineers with a thin grasp of physics. We need both in one person. Scientific computing skills sit in academia and national research groups, and many of those people have never shipped to a company. The overlap is small, and good people are often reluctant to leave research.

Where to find them

Look to physics, applied mathematics, astronomy and computational science groups where people already pair simulation with learning. Common titles include Physics ML Researcher, computational scientist, research software engineer and postdoctoral researcher. They also come from scientific machine learning teams at frontier labs, engineering simulation companies, climate and weather modelling groups, and observatory data teams.

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 a Physics-Informed ML Scientist

  1. Describe a model you built that had to respect a physical constraint. How did you enforce it and how did you check it held?

    A strong answer names the constraint, the method (loss term, architecture or data) and the tests used. Look for measured violations, not assertions. Worry if constraints were only mentioned in passing or never tested.

  2. How do you decide between a soft penalty and a hard architectural constraint?

    Good answers weigh exactness against trainability, flexibility and cost, with examples of each. They know soft penalties can leave small violations that accumulate. Worry about a single default applied to every problem.

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

Job description you can post

We are building learned models that stay faithful to the physics they describe. We need a scientist who can turn simulations and observations into fast, tested approximations that researchers and product teams can rely on. You will work closely with domain experts and engineers.

What you will do

  • Build learned approximations of physical systems that incorporate known constraints
  • Train and compare surrogates against simulation codes and observational data
  • Design tests for conservation laws, symmetries and out-of-sample accuracy
  • Quantify and communicate model uncertainty
  • Work with domain scientists to define what correct behaviour means
  • Document model limits and safe operating ranges

What you will bring

  • Demonstrated work building physics-informed or surrogate models
  • Strong deep learning skills shown in shipped or published work
  • Solid grounding in a physical or mathematical science
  • Experience validating models beyond the training distribution
  • Practical uncertainty quantification experience
  • Clear written communication of technical findings

Nice to have

  • Experience with formal verification or theorem provers
  • Familiarity with observatory or large simulation datasets
  • Experience moving research models into production use

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

Skills that matter

  • Scientific domain expertiseAdvancedApply disciplinary theory and experimental standards
  • Formal verificationAdvancedProve stated properties under explicit assumptions
  • Deep learningWorkingBuild and diagnose trainable neural models
  • Uncertainty quantificationWorkingEstimate confidence, calibration and model uncertainty

Other roles in aI for physics, mathematics and astronomy

EmergingAI Mathematical Discovery ResearcherAn AI Mathematical Discovery Researcher uses learning systems and formal tools to propose conjectures and algorithms, then gets them independently…Hiring nowAstronomy AI ScientistAn Astronomy AI Scientist analyses astronomical observations with calibrated statistical learning. They give a company catalogues and detection…EmergingQuantum Machine Learning ResearcherA Quantum Machine Learning Researcher tests whether quantum methods can beat classical ones for learning problems under realistic limits. They save…

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.

← AI for physics, mathematics and astronomyAll new job families

Start a search

Hiring a Physics-Informed ML Scientist?

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.