AI labor research
Remote AI Hiring Trends: Roles, Signals, and Limits
AI hiring is broader than the words “AI engineer.” This report maps visible AI and machine-learning signals across active remote roles and explains how candidates can distinguish research, applied engineering, data, platform, and product work.
live dataset · September 11, 2026
138 roles with explicit AI or ML signals
This narrow subset is 8% of the active inventory. It uses title/tag matches and may omit AI work described only in the posting body.
| Normalized category | Matching roles | Share of AI/ML subset |
|---|---|---|
| Data & AI | 123 | 89% |
| DevOps & Infra | 9 | 7% |
| Marketing & Growth | 4 | 3% |
| Software Engineering | 2 | 1% |
Defining an AI-related listing
The live snapshot flags titles and tags that contain explicit signals such as artificial intelligence, machine learning, ML, LLM, data scientist, research scientist, or applied scientist. This is intentionally narrower than guessing from company branding. It produces an auditable subset, although it will miss roles whose AI work is described only in the body.
A match does not prove that model development is the main responsibility. Product managers, infrastructure engineers, security specialists, and solutions engineers may support AI products without training models. The category mix in the snapshot is therefore as important as the total count.
Five different kinds of AI work
Research roles focus on new methods and rigorous evaluation. Applied ML roles adapt models to product constraints. Data and ML platform roles build pipelines, feature systems, serving, and observability. Product engineering roles integrate models into reliable user experiences. Product and go-to-market roles decide where the technology produces customer value.
These families demand different evidence. A research publication may matter for a scientist but not replace production ownership for an ML platform engineer. An LLM demo may show initiative but does not demonstrate evaluation, latency control, privacy, cost management, or incident response unless those constraints are documented.
Use broad occupational data carefully
The U.S. Bureau of Labor Statistics does not publish one occupation called remote AI engineer. Its Data Scientists profile and Software Developers profile provide broader context for tasks, education, pay, and projected employment. Those estimates cover on-site and remote work and should not be presented as a forecast for this board's AI subset.
BLS attributes software demand partly to continued expansion of AI, Internet of Things, robotics, and automation applications. That supports the claim that AI affects broader software work; it does not support a claim that every role with an AI label will grow at the same rate.
A credible portfolio for applied AI
Start with a real task and a measurable baseline. Document the dataset or retrieval source, evaluation method, failure cases, privacy constraints, latency, and cost. If a simpler deterministic system performs well, say so. Employers need judgment about when to use a model, not only evidence that you can call an API.
For remote collaboration, add a short design memo and an experiment log. Explain what you tried, why a result changed, and which risks remain. This lets a reviewer assess reasoning asynchronously and separates a maintained engineering project from a polished but unexamined demo.
- Name the role family: research, applied ML, platform, product engineering, or product.
- Match the portfolio artifact to that family's core responsibility.
- Report evaluation and failure modes alongside the successful example.
- Track fresh openings because labels and teams change quickly.
Limits of trend language
A current snapshot can show concentration, not direction. Calling something a trend normally requires consistent observations over several periods. Republished requisitions, changed tags, and additions to the company registry can all change counts even when economy-wide demand is stable.
We therefore present the current signal and its method rather than a growth headline. The monthly hiring report supplies the denominator, while the skills report helps candidates see which adjacent technologies appear in the same structured metadata.
Sources and further reading
continue with live listings
Browse relevant remote jobs
related research