AI-title intelligence, cited

AI job titles, with the data that’s actually sourceable

New AI titles — AI Engineer, ML Engineer, MLOps, Prompt Engineer — mostly have no government occupation, so no one can honestly quote “the” salary. Instead of inventing a number, RoleMath shows three sourced things per title: the nearest BLS occupation for cited pay and outlook (labeled a proxy), what employers actually ask for from public job postings, and how AI is being usedin that kind of work — each dated, cited, and clear about what it can’t say.

The method, in one paragraph

How we handle a title with no occupation code

Three layers, never blended. Pay/outlookis occupation-level BLS for the nearest established occupation, shown as a proxy — never a title-level or AI-caused figure. Hiring signalis a dated sample of public job postings (Greenhouse, Lever, Ashby, Workday) — the skills, certs, and companies, never a demand count or a scraped salary. AI usageis the Anthropic Economic Index augment-vs-automate split — descriptive of how AI is used, never an automation-risk score. Where a title can’t be anchored, we say so.

The titles — with the full cited picture

AI titles, mapped to sourceable data

AI Engineer

Occupation-anchored

Also seen as: Applied AI Engineer, LLM Engineer, GenAI Engineer

Nearest occupation — pay & outlook (proxy) · medium confidence
Occupation
Software Developers (15-1252)
Occupation median pay
$135,980
Growth 2024–34
+15.8%
Openings / yr
~115,200
Typical entry
Bachelor's degree

Primarily software-engineering work integrating ML models and LLMs into applications; the employer signal (Python, AWS, API, LLM) straddles Software Developers and Data Scientists, with Software Developers the nearest single occupation.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
39.2% augmentation (worked through with AI)60.8% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

AI Security Engineer

Occupation-anchored

Also seen as: ML Security Engineer, AI Red Team Engineer

Nearest occupation — pay & outlook (proxy) · medium confidence
Occupation
Information Security Analysts (15-1212)
Occupation median pay
$129,180
Growth 2024–34
+28.5%
Openings / yr
~16,000
Typical entry
Bachelor's degree

Security engineering applied to AI/ML systems; nearest occupation is Information Security Analysts.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
IAM · 74AWS · 46Python · 43Cybersecurity · 40Azure · 39GCP · 34vulnerability management · 30Kubernetes · 27Terraform · 24Problem solving · 23

Certifications named: Security+ (16), CCNA (9), PMP (2), CySA+ (1), Network+ (1)

Example companies in the pool (alphabetical): 1Password, Affirm, Asana, Automox, Bishop Fox, Bitwarden, Booz Allen Hamilton, Box

Based on a sample of 108 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
23.9% augmentation (worked through with AI)76.1% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

AI Solutions Architect

Occupation-anchored

Also seen as: AI Architect, ML Solutions Architect

Nearest occupation — pay & outlook (proxy) · low confidence
Occupation
Software Developers (15-1252)
Occupation median pay
$135,980
Growth 2024–34
+15.8%
Openings / yr
~115,200
Typical entry
Bachelor's degree

Architecture-level design of AI/ML systems; the work straddles Software Developers and Computer Systems Engineers/Architects with no clean single occupation, so the anchor is low confidence.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
39.2% augmentation (worked through with AI)60.8% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

AI Specialist

Occupation-anchored

Also seen as: AI/ML Specialist, AI Generalist

Nearest occupation — pay & outlook (proxy) · medium confidence
Occupation
Data Scientists (15-2051)
Occupation median pay
$120,230
Growth 2024–34
+33.5%
Openings / yr
~23,400
Typical entry
Bachelor's degree

RoleMath role-spine umbrella for applied AI work; the Data Scientists anchor is provisional and senior-skewed, so entry pay sits well below the occupation median.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
52.6% augmentation (worked through with AI)47.4% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

Computer Vision Engineer

Occupation-anchored

Also seen as: CV Engineer, Computer Vision Scientist

Nearest occupation — pay & outlook (proxy) · medium confidence
Occupation
Data Scientists (15-2051)
Occupation median pay
$120,230
Growth 2024–34
+33.5%
Openings / yr
~23,400
Typical entry
Bachelor's degree

Specializes in image and video machine learning; BLS folds this work into Data Scientists, and public postings are senior and research-skewed, so entry pay sits below the occupation median.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
52.6% augmentation (worked through with AI)47.4% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

Data Engineer (AI/ML)

Occupation-anchored

Also seen as: ML Data Engineer, AI Data Engineer

Nearest occupation — pay & outlook (proxy) · medium confidence
Occupation
Software Developers (15-1252)
Occupation median pay
$135,980
Growth 2024–34
+15.8%
Openings / yr
~115,200
Typical entry
Bachelor's degree

Builds the data pipelines that feed ML systems; nearest occupation is Software Developers (Database Architects is an alternate).

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
39.2% augmentation (worked through with AI)60.8% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

MLOps Engineer

Occupation-anchored

Also seen as: ML Platform Engineer, ML Infrastructure Engineer

Nearest occupation — pay & outlook (proxy) · low confidence
Occupation
Software Developers (15-1252)
Occupation median pay
$135,980
Growth 2024–34
+15.8%
Openings / yr
~115,200
Typical entry
Bachelor's degree

Infrastructure and DevOps work for ML systems; straddles Software Developers and Network/Systems Administrators with a thin standalone public-posting sample, so the anchor is low confidence.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
39.2% augmentation (worked through with AI)60.8% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

Machine Learning Engineer

Occupation-anchored

Also seen as: ML Engineer, MLE

Nearest occupation — pay & outlook (proxy) · medium confidence
Occupation
Data Scientists (15-2051)
Occupation median pay
$120,230
Growth 2024–34
+33.5%
Openings / yr
~23,400
Typical entry
Bachelor's degree

BLS folds machine-learning modeling work into Data Scientists; public postings are senior and research-skewed, so entry pay sits below the occupation median.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
52.6% augmentation (worked through with AI)47.4% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

NLP Engineer

Occupation-anchored

Also seen as: Natural Language Processing Engineer

Nearest occupation — pay & outlook (proxy) · medium confidence
Occupation
Data Scientists (15-2051)
Occupation median pay
$120,230
Growth 2024–34
+33.5%
Openings / yr
~23,400
Typical entry
Bachelor's degree

Specializes in language and large-language-model machine learning; BLS folds this work into Data Scientists; the sample is senior-skewed.

This is the nearest established occupation, shown as a proxy — pay is set by the occupation and location, not by the job title or by AI, and entry roles sit below the median. Median: BLS OEWS 2025-05. Growth/openings/education: BLS Employment Projections 2024-2034.

See the full cited role page →

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
52.6% augmentation (worked through with AI)47.4% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

AI Product Manager

Hiring signal only

Also seen as: AI PM, GenAI Product Manager

No single occupation — candidate matches

No single clean federal occupation — the work straddles Project Management Specialists (13-1082), management, and computer occupations. RoleMath withholds a single pay anchor and shows the employer hiring signal instead.

Candidate occupations: Project Management Specialists (13-1082), Computer & Information Systems Managers (11-3021), Data Scientists (15-2051). We withhold a single pay figure rather than pick one.

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

AI Research Scientist

Hiring signal only

Also seen as: Research Scientist (AI/ML), Member of Technical Staff (Research)

What employers ask for
Machine learning · 451Python · 392LLM · 293AWS · 135SQL · 129PyTorch · 127OpenAI · 110Problem solving · 109Okta · 108API · 104

Certifications named: nonein this sample — for these roles employers list skills, not certs. (A real signal for AI/ML work; it differs sharply from, say, network roles, where CCNA is named often.)

Example companies in the pool (alphabetical): 1Password, Affirm, Airbnb, Anthropic, Asana, Aviatrix, BeyondTrust, Booz Allen Hamilton

Pooled across AI/ML job postings: there is no title-specific posting sample for this title yet, so this is the nearest matching hiring pool.

Based on a sample of 753 public job postings (June 2026). Qualitative employer-language sample only; do not use as official demand, market-size, salary, or certification ROI evidence. This is a sample of public postings, not representative of the whole market.

How AI is used in this work
42.1% augmentation (worked through with AI)57.9% automation (delegated to AI)

Measured usage of one AI tool (Claude conversations mapped to this occupation's O*NET tasks), May 2026, from the Anthropic Economic Index (CC-BY). 'Augmentation' means a person worked through a task with AI; 'automation' means the task was delegated. This is descriptive of how AI is used, not job loss, not automation risk, not an AI-exposure score, and not a prediction. A higher automation share does not mean a role is 'more at risk.'

Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation; most tech desk-roles skew toward automation-style (delegation) usage, which is a fact about how people prompt the tool for those tasks, not a statement that those jobs are more automatable.

Source: Anthropic Economic Index (release 2026-06-26), data licensed CC-BY, via Hugging Face Anthropic/EconomicIndex. Massenkoff, Lyubich, Sacher, Hitzig, Zhang, Heller & McCrory (2026), Anthropic Economic Index report: Cadences.

Where we won’t guess

Emerging titles we won’t put a number on

AI Ethicist

Emerging title

Also seen as: Responsible AI Specialist, AI Policy Analyst

Why no pay figure

No durable federal occupation and too small a public-posting sample to characterize honestly.

This is an emerging title with no dedicated government occupation. Rather than attach a job-board number with no sample size or date, we say so — a figure we cannot source honestly is one we don’t publish.

Generative AI Designer

Emerging title

Also seen as: AI UX Designer, GenAI Designer

Why no pay figure

Design-adjacent work with no durable federal occupation and a thin sample; RoleMath does not attach a pay figure.

This is an emerging title with no dedicated government occupation. Rather than attach a job-board number with no sample size or date, we say so — a figure we cannot source honestly is one we don’t publish.

Prompt Engineer

Emerging title

Also seen as: LLM Prompt Engineer

Why no pay figure

No durable federal occupation; prompt engineering appears as a skill line-item inside AI/ML postings (78 of the AI-specialist sample), not yet a standalone occupation with a characterizable sample.

This is an emerging title with no dedicated government occupation. Rather than attach a job-board number with no sample size or date, we say so — a figure we cannot source honestly is one we don’t publish.

Common questions

AI job titles, answered honestly

Why don’t AI job titles have their own salary?
The U.S. federal occupation system (BLS SOC / O*NET) updates on a multi-year cadence and has no dedicated code for most AI titles like AI Engineer or ML Engineer. There is no official "AI Engineer salary," and a single number scraped from a job board has no sample size, date, or outlook behind it. RoleMath anchors each AI title to the nearest established occupation for cited pay and outlook, labels it a proxy, and separately shows what employers actually ask for in public job postings.
Where does the “what employers ask for” data come from?
From public applicant-tracking job boards (Greenhouse, Lever, Ashby, Workday) only — the skills, tools, certifications, and title terms employers name in real postings, plus a sample of companies hiring. It is a dated, sampled hiring signal, not total market demand, not a salary, and not a return-on-investment figure. RoleMath never scrapes job-board salary or posts demand counts.
Does a higher “automation” share mean the job is at risk?
No. The augmentation-versus-automation split comes from the Anthropic Economic Index and is descriptive of how one AI tool is actually used in that kind of work — "augmentation" means working through a task with AI, "automation" means delegating it. It is not job loss, not automation risk, and not a prediction. Across all occupations the May 2026 split was 51.4% augmentation / 48.6% automation.
Do you need a certification to get an AI job?
In our sample of 753 public AI/ML job postings (Greenhouse, Lever, Ashby, Workday), employers named zero certifications — they listed skills like machine learning, Python, LLMs, and PyTorch instead. That contrasts sharply with traditional IT roles in the same sample, where certs are named often (network-administrator postings named CCNA 42 times, project-coordinator postings named PMP 39 times). This is a dated, sampled hiring signal, not a rule — but for AI/ML roles specifically, demonstrable skills appear to matter more to employers than a certification. The AI-security crossover is the exception: those postings do name Security+ and CCNA.
Which AI titles can’t you anchor at all?
Prompt Engineer, AI Ethicist, and Generative AI Designer have no durable federal occupation and too small a public-posting sample to characterize honestly, so we mark them emerging and explain why instead of inventing a number. Prompt engineering, for example, shows up as a skill line-item inside AI/ML postings, not yet as a standalone occupation.

Build a cited plan for the role you want

RoleMath maps AI roles to cited occupation pay and outlook, sells nothing, and isn’t influenced by who pays us. See the pay-and-outlook view or build a plan for your background and target role.