Hiring nowLearning theory and statistical machine learning

Automated Machine Learning Researcher

Also called AutoML Scientist

An Automated Machine Learning Researcher builds search procedures that choose and tune models within a fixed compute budget. It matters because teams get better models faster, with results they can trust on data the search never saw.

What they do

They design the search that picks models, features and hyperparameters, and they decide what it is allowed to spend. Day to day they write search code in Python, run experiments across many datasets, and compare methods against simple baselines such as random search. They set out the benchmark protocol: which datasets, which splits, how the budget is counted. They hunt for leakage, where information from validation or test data seeps into the search. They read results sceptically and report where a method fails as well as where it wins.

What a strong hire delivers

In the first months, a strong hire delivers a working search procedure that beats the team's current manual tuning at the same compute cost. They also deliver a written benchmark protocol that others can rerun, with clean separation between data used to search and data used to judge. Expect a clear account of where the search helps, where it does not, and what budget makes it worthwhile.

How you can tell they are good

Ask them to walk through a search method they built and how they proved it worked on data it never touched. Good candidates volunteer the baselines they beat, the budget they measured against and the leakage they found. Ask what they would do if the gains vanished on a new dataset. Worry about anyone who only quotes the best score.

Why it is hard to fill

Titles are inconsistent. The work appears under AutoML, hyperparameter optimisation, experimental design or plain research scientist. The skill sits between statistics and engineering, and many people have one without the other. Strong candidates often sit in research teams and are not looking. Plenty have tuned models, but far fewer have designed a leakage-safe evaluation of the search itself.

Where to find them

Look in research groups working on hyperparameter optimisation, Bayesian optimisation and meta-learning, and in applied science teams at model providers and large technology companies. Statisticians and mathematicians who moved into machine learning are a good source. So are engineers from platform teams who built internal tuning systems. Academic labs and PhD graduates with open benchmark work are worth a look.

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 Automated Machine Learning Researcher

  1. Describe a search procedure you built. What budget did it run under and what did you compare it against?

    A strong answer names the search method, the compute budget and honest baselines such as random search. They explain how the budget was counted. Worry if there is no baseline or the comparison used unequal compute.

  2. How do you stop information leaking from validation or test data into a model search?

    They separate data used for search from data used for judgement, handle preprocessing inside the folds and keep a final untouched set. They mention repeated peeking. Worry if they treat leakage as only a train-test split issue.

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

Job description you can post

We are hiring an Automated Machine Learning Researcher to build the systems that choose and tune our models. The aim is better models for less compute, judged on data the search has never seen. You will work closely with our research and engineering teams.

What you will do

  • Design and build search procedures for model selection and hyperparameter optimisation
  • Set and enforce compute budgets for every search
  • Write and maintain a benchmark protocol others can rerun
  • Find and remove sources of leakage in our evaluation
  • Compare new methods against simple baselines at matched cost
  • Report results honestly, including where methods fail

What you will bring

  • Demonstrated work on hyperparameter optimisation or model search
  • Strong grounding in probability, statistics and optimisation
  • Experience designing experiments that hold up on untouched data
  • Fluent Python, with scikit-learn and PyTorch
  • Clear written reporting of results and limitations

Nice to have

  • Experience with Bayesian optimisation or meta-learning
  • Published or open benchmark work
  • Experience building internal tuning platforms

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

Skills that matter

  • Probability and statisticsAdvancedQuantify variability and statistical evidence
  • Linear algebra and optimisationAdvancedReason about matrix operations and objective functions
  • Experimental designWorkingChoose comparisons that identify credible effects
  • Deep learningWorkingBuild and diagnose trainable neural models

Other roles in learning theory and statistical machine learning

Hiring nowMachine Learning Research ScientistA Machine Learning Research Scientist develops and tests learning methods against explicit hypotheses. They give the company evidence about what…Hiring nowStatistical Learning ScientistA Statistical Learning Scientist builds prediction methods that show their assumptions and their uncertainty. For a company shipping models into…Hiring nowLearning Theory ResearcherA Learning Theory Researcher works out what learning algorithms can and cannot do, and proves it. For an AI company, that means firmer ground for…

Related roles in other families

Hiring nowPretraining Research EngineerFoundation model pretrainingA pretraining research engineer runs controlled experiments on how large models learn from broad data. The work decides which architectures, data…Hiring nowModel Architecture ResearcherFoundation model pretrainingA Model Architecture Researcher tests changes to how a model is built, to improve its behaviour or efficiency. The work tells a company which…EmergingScaling Science ResearcherFoundation model pretrainingA Scaling Science Researcher works out how model performance changes with compute, data and model size. Their estimates tell a company how to spend…EmergingTraining Dynamics ResearcherFoundation model pretrainingA Training Dynamics Researcher works out why large training runs wobble, stall or diverge, and fixes the cause. Fewer failed runs means less wasted…

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.

← Learning theory and statistical machine learningAll new job families

Start a search

Hiring an Automated Machine Learning 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.