Loading the job postings and counting them…
Loading the job postings and counting them…
AI Career Compass · role
Develops new methods and studies how models behave. Model capabilities, training methods, interpretability, safety and science applications.
Measured In our tracked sample, 55 of 1,694 AI postings are for research scientists, about 3%.1
Measured The skills that research scientist postings ask for most are machine learning, in 37 of 55, and the programming language Python, in 19 of 55.12
Written by an AI analyst (claude-opus-5-5) and checked by code against the numbers it cites.
What these postings say the work involves, counted from their lists of duties. A posting can list several.
Out of these 55 postings. “Asked for” adds required, one of several options, and listed; preferred is counted apart.
Programming, data, maths and statistics, and domain knowledge
| Skill | Asked for | Required | One of several | Listed | Preferred |
|---|---|---|---|---|---|
| Python | 19 of 55 | 5 | 3 | 11 | 1 |
| Software engineering practice | 11 of 55 | 2 | 0 | 9 | 2 |
| Science (biology, chemistry, physics) | 5 of 55 | 3 | 0 | 2 | 1 |
| Data analysis | 3 of 55 | 0 | 0 | 3 | 2 |
| TypeScript or JavaScript | 3 of 55 | 0 | 2 | 1 | 0 |
| Go | 2 of 55 | 0 | 2 | 0 | 0 |
| Rust | 2 of 55 | 0 | 2 | 0 | 0 |
| Statistics and probability | 2 of 55 | 0 | 0 | 2 | 0 |
Modelling, training, retrieval, agents, evaluation and other methods
| Skill | Asked for | Required | One of several | Listed | Preferred |
|---|---|---|---|---|---|
| Machine learning | 37 of 55 | 5 | 0 | 32 | 4 |
| PyTorch | 16 of 55 | 3 | 1 | 12 | 3 |
| Large language models | 15 of 55 | 2 | 1 | 12 | 4 |
| Deep learning | 13 of 55 | 2 | 0 | 11 | 5 |
| Reinforcement learning | 12 of 55 | 1 | 2 | 9 | 7 |
| Fine-tuning and post-training | 11 of 55 | 3 | 0 | 8 | 4 |
| Alignment and AI safety research | 6 of 55 | 0 | 0 | 6 | 0 |
| JAX | 6 of 55 | 1 | 1 | 4 | 0 |
Deployment, infrastructure, reliability, security, monitoring and cost
| Skill | Asked for | Required | One of several | Listed | Preferred |
|---|---|---|---|---|---|
| Distributed training | 10 of 55 | 1 | 0 | 9 | 1 |
| Security engineering | 3 of 55 | 0 | 0 | 3 | 0 |
| Data pipelines (Spark, Airflow, Kafka) | 2 of 55 | 0 | 0 | 2 | 4 |
| MLOps and ML platforms | 2 of 55 | 1 | 0 | 1 | 3 |
| System design and architecture | 2 of 55 | 0 | 0 | 2 | 1 |
| AI security (prompt injection, model abuse) | 2 of 55 | 0 | 0 | 2 | 1 |
| Inference speed and efficiency | 1 of 55 | 0 | 0 | 1 | 4 |
| Cloud platforms (AWS, GCP, Azure) | 1 of 55 | 0 | 0 | 1 | 2 |
Publishing and leading research
| Skill | Asked for | Required | One of several | Listed | Preferred |
|---|---|---|---|---|---|
| Leading research | 16 of 55 | 4 | 0 | 12 | 3 |
| Research publications | 15 of 55 | 3 | 0 | 12 | 9 |
| Open-source contributions | 1 of 55 | 0 | 0 | 1 | 4 |
Customers, communication, product thinking and leadership
| Skill | Asked for | Required | One of several | Listed | Preferred |
|---|---|---|---|---|---|
| Communication | 12 of 55 | 2 | 0 | 10 | 3 |
| Technical writing | 4 of 55 | 1 | 0 | 3 | 0 |
| Leading and mentoring | 3 of 55 | 1 | 0 | 2 | 2 |
| Product thinking | 3 of 55 | 2 | 0 | 1 | 0 |
| Working with customers | 0 of 55 | 0 | 0 | 0 | 2 |
AI rules, policy and trust and safety
| Skill | Asked for | Required | One of several | Listed | Preferred |
|---|---|---|---|---|---|
| Risk and compliance | 3 of 55 | 0 | 0 | 3 | 0 |
Measured Research publications are listed as a preferred qualification in 9 of 55 research scientist postings we track.1
Measured Reinforcement learning, a way to train models by rewarding good results, is listed as a preferred qualification in 7 of 55 of these postings.1
Employers list these as preferred. That they ask does not show that these skills decide who is hired.
The systems these postings expect people to build, run or work within.
Only what the postings state. Most do not say everything, so each number has its own count of postings.
13 of 55 postings state a minimum; the median is 3 years.
23 of 55 postings state a degree; 13 of them accept equivalent experience instead.
3 more postings list a PhD as preferred.
12 postings are remote. Of these, 12 name where you must live or work from, 0 do not say, and 0 say anywhere.
What employers advertise, grouped only when the ranges are comparable. Not a prediction of what you would earn.
Measured For full-time jobs in the United States, the median of advertised range midpoints is 397,500 US dollars per year in base pay, from 18 comparable pay disclosures in the postings we track.12
48 of 55 postings (87%) advertise a pay range. Employers that publish ranges may differ from those that do not, and some places require it by law, so these ranges do not describe all pay.
| Group of alike ranges | Postings | Median of range midpoints | Middle half of midpoints |
|---|---|---|---|
| USD, a year, base pay, United States, full-time5 employers | 18 | $397,500 | $311,875 to $440,000 |
This is advertised pay, not what anyone earns. We pool ranges only when currency, pay period, pay type, country and employment type are the same, and only with at least 10 of them. Nothing is converted. 30 other ranges could not be pooled (period not stated, several countries, or too few alike).
Advice built from what these 55 postings ask for. Tick what you can already do; we suggest up to three steps and one portfolio project, each with the requirement behind it and a way to show the result.
Analysis Tracking began on 6 October 2026, so this window holds one reading so far and shows no trend. Many open listings are not new: only 10 of 55 were first published in the 30 days up to the data cutoff.1
Analysis The 55 research scientist postings we track come from 18 employers that publish their jobs through public job feeds. This is not the whole job market, and some of the largest employers are not included.12
Analysis Some listings stay open a long time: 12 of 55 were first published more than a year before the data cutoff.1
How everything is counted: our method. All roles: AI Career Compass.