What AI jobs are in demand: 18 job titles ranked by open roles

What AI jobs are in demand is a question with a measurable answer, if you are willing to accept open postings as the measure. AI Jobs Map tracks 18 AI job families, and on September 17, 2026 it listed 28,283 AI roles posted in the last 30 days, of which 9,529 were posted in the last 7 days. This reference ranks every title by open roles, gives the remote and entry level share of each, and says in one paragraph what each title means.

One caveat before the table. Postings measure hiring intent, not headcount: a company that reposts a role every month counts every month. The figures are still checkable, because every listing links to the posting it came from.

The 18 AI job titles, ranked

The first row is not a title; it is every posting that matches any AI title, and it is smaller than the sum of the rows below it because one posting can match several. The top three titles on September 17, 2026 were AI Engineer, Data Scientist and Machine Learning Engineer, and they held 85% of the total between them. Three titles are most of the market; the other fifteen are where the specialists are.

How fast the market moves

9,529 of the 28,283 open roles were posted in the previous 7 days, which is 34% of a 30-day window. The last three days alone held 4,931 roles; if you check one page daily, that is the one.

7,086 of the open roles stated a salary in the posting, a 25% share. AI Jobs Map shows a salary on a job page only when the employer stated one, and never publishes an estimate as if it were an offer, so the pages carry no figure for the rest. The section on pay below says what that means for the “high paying AI jobs” question.

The titles, one by one

Each title is described with the wording from its own page, then its remote and entry level share and one observation.

AI Engineer Jobs, 10,474 open roles. AI engineer roles building LLM-powered products, agents, retrieval systems and AI platform infrastructure. 13% remote, 9% entry level. The largest title on the site and the one most postings default to when the work is building on a model API.

Data Scientist Jobs, 8,778 open roles. Data scientist roles across analytics, experimentation and applied machine learning. 10% remote, 6% entry level. The broadest title here; many of these postings are analytics roles with no machine learning in them.

Machine Learning Engineer Jobs, 4,829 open roles. Machine learning engineer roles: model training, feature pipelines, evaluation and production ML systems. 15% remote, 9% entry level. The title employers use when the model is trained in-house rather than bought.

Agentic AI Jobs, 4,111 open roles. Agentic AI roles: engineers and researchers building autonomous agents, tool use and multi-step AI workflows. 13% remote, 5% entry level. New as a title and already larger than generative AI; the postings are application work, not model training.

Generative AI Jobs, 3,339 open roles. Generative AI roles across engineering, research and product. 8% remote, 6% entry level. A label that spans engineering, research and product; read the posting for which one it is.

Forward Deployed Engineer Jobs, 2,891 open roles. Forward deployed engineer (FDE) roles: engineers embedded with customers to ship AI systems into production. 12% remote, 6% entry level. A title that barely existed outside one company a few years ago, now a job family of its own.

AI Product Manager Jobs, 1,646 open roles. AI product manager roles shipping AI features, agents and ML-driven products. 20% remote, 11% entry level. The non-engineering title with the most postings; most ask for prior product management, not prior AI work.

AI Automation Jobs, 1,251 open roles. AI automation roles: building agent workflows and LLM-driven automation for operations, support and sales. 19% remote, 11% entry level. Matched on the title plus an LLM term in the description, so it captures operations and RevOps roles that build with agents.

Applied Scientist Jobs, 1,063 open roles. Applied scientist roles at the boundary of ML research and production. 3% remote, 1% entry level. Concentrated in a few large employers and rarely remote.

AI Research Jobs, 974 open roles. AI research scientist and research engineer roles in industry labs and applied research teams. 14% remote, 8% entry level. The highest bar of any title here and the one that most often asks for publications.

MLOps Engineer Jobs, 873 open roles. MLOps and ML platform roles: training infrastructure, model serving, monitoring and pipelines. 13% remote, 5% entry level. Postings read like platform engineering with a model in the pipeline.

Responsible AI Jobs, 609 open roles. Responsible AI, AI governance, AI safety and AI policy roles. 12% remote, 3% entry level. The smallest of the non-engineering titles, concentrated in regulated industries.

Computer Vision Jobs, 593 open roles. Computer vision engineer and scientist roles: perception, imaging, video and multimodal models. 9% remote, 12% entry level. A high entry level share, because much of the work sits in robotics and manufacturing.

Deep Learning Jobs, 290 open roles. Deep learning engineer and researcher roles. 6% remote, 9% entry level. Mostly research and hardware-adjacent roles; used less each year as it becomes the default.

LLM Engineer Jobs, 259 open roles. LLM engineer roles: fine-tuning, inference, evaluation, prompt systems and large language model applications. 22% remote, 13% entry level. Small as a title, large as a skill; most of this work is posted under AI engineer.

AI Prompt Engineer Jobs, 139 open roles. Prompt engineer roles: prompt design, evaluation and LLM behaviour tuning. 14% remote, 24% entry level. The smallest engineering title on the site and the one with the highest entry level share.

NLP Jobs, 40 open roles. NLP engineer and scientist roles: language models, information extraction, speech and text systems. 15% remote, 10% entry level. Nearly absorbed into LLM engineering; the postings that remain are speech, extraction and search.

High paying AI jobs

This site does not rank titles by pay, and it is worth saying why rather than leaving the section out. Only 25% of postings state a salary, the ones that do are skewed towards jurisdictions that require it, and a median from that subset would describe the disclosure law more than the job. Job pages show the stated range when there is one; that is as far as the data honestly goes.