Hiring nowContinual, adaptive and federated learning

Online Learning Scientist

Also called Adaptive ML Scientist

An Online Learning Scientist builds models that keep learning from arriving data under explicit feedback rules. It matters because products that stand still decay as users, markets and data sources change.

What they do

Designs and runs learners that update from a stream of observations, rather than from a fixed training set. Decides what feedback counts, how fast the model should move, and what it must never forget. Builds replay setups so changes can be tested on past data before they touch production. Sets up drift tests and alerts. Works with engineers on the data path, and with privacy colleagues where data stays with its owner. Writes down the update rules so others can audit them. Reviews failures, such as a model that chases noise or freezes too early, and fixes the cause.

What a strong hire delivers

Within the first months, a working online learner on one live problem, with the feedback rules written down and agreed. A replay evaluation that the team trusts and reruns. Drift tests with clear thresholds and a named response. A short report on regret, stability and how quickly the model recovers after a shift, so decisions about further investment rest on evidence.

How you can tell they are good

Ask them to walk through a learner they updated from live data and what went wrong. Strong people explain the feedback rule, the stability trade-off and how they tested offline before release. They can say how they would know the model was drifting. Be wary of anyone who talks only about retraining on a schedule or quotes offline accuracy alone.

Why it is hard to fill

The title is rarely used, so searches on it miss most of the people. The skills sit across research, optimisation and privacy engineering, and few have run a learner in production. Many strong candidates are known as continual learning researchers, bandit specialists or federated learning engineers. Candidates who can show both the theory and a live system are scarce, and they are often not looking.

Where to find them

Look at ML researchers with a sequential or streaming focus, distributed optimisation specialists, and privacy engineers who have worked on federated setups. Recommendation, ads, fraud and personalisation teams often hold people who have run live learners. Research groups at universities and frontier labs also produce them. Adjacent titles include applied scientist, bandit researcher and ML platform engineer.

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 Online Learning Scientist

  1. Describe a model you updated from arriving data. What were the feedback rules and why?

    Strong answers name the signal, its delay and its bias, and explain why the update rule suited it. They show awareness of feedback loops. Worry if feedback is treated as clean ground truth or the rules were never written down.

  2. How would you tell drift in the data from a fault in the pipeline?

    Good answers check upstream data, schema and volume before blaming the world. They mention statistical drift tests, comparison against reference windows and a staged response. Worry if the answer is simply to retrain.

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

Job description you can post

We build products that learn from live data as it arrives. We need an Online Learning Scientist to design learners that adapt under clear feedback rules, and to prove they stay stable as conditions change. You will work with engineers, researchers and privacy colleagues.

What you will do

  • Design and ship online learners that update from streaming observations
  • Define and document feedback rules and update policies
  • Build replay and production evaluations for regret, stability and drift response
  • Set up drift tests, thresholds and response plans
  • Investigate and fix failures in live learning systems
  • Work with engineers on data paths and with privacy colleagues on distributed ownership

What you will bring

  • Demonstrated work on streaming, online or continual learning systems
  • Experience evaluating models on time-ordered data
  • Working knowledge of distributed systems and optimisation
  • Clear thinking about uncertainty and how to quantify it
  • Ability to explain trade-offs to non-specialists
  • Evidence of a learner that ran in production or a realistic replay

Nice to have

  • Experience with federated learning frameworks
  • Privacy engineering background
  • Published or shared work on sequential benchmarks

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

Skills that matter

  • Continual learningAdvancedAdapt models while measuring retention and drift
  • Privacy engineeringAdvancedLimit disclosure and unauthorised personal-data use
  • Distributed systemsWorkingDesign reliable work across networked machines
  • Uncertainty quantificationWorkingEstimate confidence, calibration and model uncertainty

Other roles in continual, adaptive and federated learning

EmergingContinual Learning ResearcherA Continual Learning Researcher builds ways to update models on new data without losing what they already do well. It matters because retraining…Hiring nowFederated Learning EngineerA Federated Learning Engineer trains models across several data holders without pooling the raw data. That lets a company use data it cannot…EmergingMeta-Learning ResearcherA Meta-Learning Researcher studies how models carry learning from one task to the next, so they adapt quickly to new tasks with little evidence…

Related roles in other families

Hiring nowComputer Vision EngineerComputer vision and spatial perceptionA 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 ResearcherComputer vision and spatial perceptionA 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 ScientistComputer vision and spatial perceptionA Video Understanding Scientist builds models that recognise actions, events and the dependencies between them in video. Companies hire for this…EmergingOpen-World Perception ResearcherComputer vision and spatial perceptionAn Open-World Perception Researcher builds vision models that know when they are looking at something unfamiliar. That matters because a system that…

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.

← Continual, adaptive and federated learningAll new job families

Start a search

Hiring an Online Learning 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.