EmergingContinual, adaptive and federated learning

Continual Learning Researcher

Also called Lifelong Learning Scientist · Continual Learning Engineer · Online Learning Engineer

A Continual Learning Researcher builds ways to update models on new data without losing what they already do well. It matters because retraining from scratch is slow and costly, and a model that forgets quietly breaks the product.

What they do

Takes a model that must keep learning and works out how to update it safely. Designs sequential training setups, runs them on benchmarks and on the company's own data, and measures what the model gains and what it loses. Tests for drift, builds retention checks into eval pipelines, and tries mitigation methods such as replay, regularisation and parameter isolation. Where data sits with different owners, works with privacy and systems engineers on federated or distributed updates. Writes up forgetting analyses that product and safety colleagues can act on.

What a strong hire delivers

Within the first months, a strong hire delivers a working sequential learning method on one real model, with a clear forgetting analysis beside it. They set up a retention test suite that runs on every update. They give the team a plain view of when an update is safe to ship, when it needs gating, and where the method still fails.

How you can tell they are good

Ask them to walk through a past update where something got worse. Good candidates name exactly which capabilities dropped, how they found out, and what they changed. Ask how they chose retention tests. Listen for honest limits, a view on trade-offs, and evidence they have run experiments, not only read the papers.

Why it is hard to fill

Very few people have done this on models at real scale, and the title is rarely used. The work often sits inside research, evaluation or safety jobs. Strong candidates tend to be in labs and are seldom looking. Many have the theory but not the production context, or the reverse. Expect to search by skill and history, not by job title. The field also moves with research breakthroughs, so the pool can shift quickly.

Where to find them

Look in frontier lab research teams, university groups working on continual or online learning, and applied ML teams that update models in production. Distributed optimisation researchers and privacy engineers who have built federated systems are a second source. Evaluation and safety gating engineers sometimes hold the retention testing skills. Titles include lifelong learning scientist, online learning engineer and research scientist.

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 Continual Learning Researcher

  1. Tell us about a model update that caused forgetting. How did you detect it and what did you do?

    A strong answer names the specific lost capabilities, the test that caught it, and the fix tried, including what did not work. Worry if they only describe aggregate scores or cannot say how forgetting was measured.

  2. How do you decide which prior capabilities to test for retention after an update?

    Good answers tie tests to product use, past failures and safety needs, and accept that coverage is never complete. Worry about a single benchmark, or a list chosen without reference to what users rely on.

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

Job description you can post

We build models that have to keep learning after release. We are hiring a Continual Learning Researcher to help us update them on new data without losing what already works. You will work with research, evaluation and safety colleagues to make updates measurable and safe to ship.

What you will do

  • Design and run sequential learning experiments on our models
  • Build and maintain retention tests that run on every update
  • Develop and compare methods to reduce catastrophic forgetting
  • Write forgetting analyses that guide release decisions
  • Set up drift tests for models in use
  • Work with systems and privacy engineers on distributed or federated updates

What you will bring

  • Shown work on continual, online or sequential learning, published or in production
  • Experience measuring capability loss, not just task gains
  • Hands-on training of large models and clear experiment records
  • Working knowledge of distributed training or federated methods
  • Ability to explain trade-offs to non-researchers

Nice to have

  • Experience with privacy engineering
  • Background in uncertainty quantification
  • Experience building safety gates for model releases

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

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…Hiring nowOnline Learning ScientistAn Online Learning Scientist builds models that keep learning from arriving data under explicit feedback rules. It matters because products that…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 a Continual 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.