Skip to content

Resume Space · Career insights

AI learning from job data

Explore job titles, skills and responsibilities from tracked job postings, see salary amounts recorded with job postings, and read research-backed ideas for where AI may help in the work.

Job data is a current snapshot across tracked locations. AI ideas are editorial suggestions, not measures of hiring demand.

Career roles

Explore job data by role

4 roles in this release

Choose a career role

Compare the current job snapshot, posted salary evidence and researched AI-use possibilities.

Showing AI Engineer job-data trends.

Current job snapshot

AI Engineer

Across tracked locations · 18 candidate job postings captured · observed Oct 28, 2026

Current snapshot
Candidate postings
18
Confirmed in-role
18
Off-role excluded
0
Awaiting role-fit review
0

Title, skill, responsibility and salary summaries use only confirmed in-role postings. Across the candidate inventory, 18 descriptions were usable; 0 were missing, 0 were truncated, and extraction failed for 0. This is a snapshot, not a forecast of job growth.

What employers mention

Job titles, skills and responsibilities

Counts show distinct job postings in this snapshot. Similar postings across locations are deduplicated.

Job titles

Among 18 postings reviewed as in-role

  1. AI Engineer#1

    2 of 18 postings · 11%

    Mentioned by 2 employers

    Small sample

  2. Staff AI Enablement Engineer#2

    1 of 18 postings · 6%

    Mentioned by 1 employers

    Small sample

  3. Senior Agentic AI engineer#3

    1 of 18 postings · 6%

    Mentioned by 1 employers

    Small sample

Showing a ranked selection from this snapshot.

Skills

Among 18 in-role postings with usable skill evidence

  1. Python#1

    15 of 18 postings · 83%

    Mentioned by 13 employers

  2. Machine learning#2

    13 of 18 postings · 72%

    Mentioned by 11 employers

  3. Model evaluation#3

    6 of 18 postings · 33%

    Mentioned by 5 employers

Responsibilities

Among 18 in-role postings with usable responsibility evidence

  1. Build AI systems#1

    8 of 18 postings · 44%

    Mentioned by 8 employers

  2. Evaluate AI models#2

    1 of 18 postings · 6%

    Mentioned by 1 employers

    Small sample

Posted salary data

Salary amounts recorded for job postings

Salary coverage uses only postings confirmed as in-role for this guide. Amounts, currencies and pay periods use the normalized values stored with job postings. The original pay context is not independently verified for every posting, so pay basis can vary. If captured compensation text clearly conflicts with a stored period, that posting is excluded. This snapshot is not a personal salary estimate; we do not convert currencies or annualize pay.

USD per year

Recorded posting values · 14 postings · 12 employers

Observed stored range

$82,100.00 – $1,066,000.00

Full span of complete ranges in this snapshot, not a typical pay range.

Median stored midpoint from 14 complete ranges: $198,150.00

14 complete ranges · 0 one-sided postings · 12 employers with complete ranges

14 of 18 in-role postings had usable stored salary fields. 4 were missing or excluded by the salary rules.

How salary postings are grouped and excluded

Groups use the stored salary minimum, maximum, currency and pay period. Not all original posting context is retained with those normalized fields, so this page does not claim every figure is base pay. When nearby captured compensation text contains both stored range endpoints and one clear, different pay period, that record is excluded. Mixed or unclear context can remain grouped under the stored period. Currency and period are kept separate; amounts are not converted or annualized.

  • 4 missing salary
  • 0 invalid salary values
  • 0 unsupported salary source
  • 0 stored periods contradicted by captured posting context
  • 0 ambiguous geographic bands
  • 0 source conflicts

Practical ideas, with human review

How AI may support this role

These are research-informed possibilities, separate from measured job demand. People remain responsible for decisions and final work.

Draft evaluation cases

For a defined model task, propose candidate normal, boundary, and adversarial examples with expected-result notes for an evaluator to review.

Possible output

A small draft evaluation set.

Human checks

  • Confirm each case represents the intended task and real input distribution.
  • Set expected outcomes independently and ask domain reviewers to label ambiguous cases.

Draft a scoring rubric

Turn an approved model-output requirement into a first-pass pass-or-fail rubric with observable criteria and failure labels.

Possible output

A draft evaluator scorecard.

Human checks

  • Make criteria observable and compare decisions with expert judgments.
  • Check for scoring bias, review disagreements, and revise the rubric before use.

Research reviewed Oct 28, 2026.

Open the full AI Engineer guide

Release seed-public-v10. AI Learning