Occupation · 2026 assessment

AI Evaluation Engineer

The AI Evaluation Engineer role combines repeatable operations with decisions that require professional context. The index measures task exposure to AI, software automation and robotics—not the probability that the occupation disappears.

Low risk level Search jobs ↗
How risk may change

A forecast estimate, not a guaranteed scenario.

22%2026
28%2028
35%2030
45%2035

Which tasks are exposed

Risk applies to individual tasks—not to a person or an entire occupation.

TaskRisk
Generating routine code and configuration47%
Preparing tests and technical documentation43%
Classifying errors and analyzing logs37%
Data migrations and routine integrations37%
Architecture and technical trade-offs16%
Security and production accountability5%

Task map

AI replaces ↔ AI augments

Each task is positioned by AI capability and the degree of human involvement still required.

  1. 1Generating routine code and configurationAI 47% · human 73%
  2. 2Preparing tests and technical documentationAI 43% · human 83%
  3. 3Classifying errors and analyzing logsAI 37% · human 74%
  4. 4Data migrations and routine integrationsAI 37% · human 75%
  5. 5Architecture and technical trade-offsAI 16% · human 79%
  6. 6Security and production accountabilityAI 5% · human 93%

Why technology can change this work

  • tools automate data preparation, experiments and parts of modelling
  • models automatically compare scenarios, prices, returns and risks
  • code, configuration and documentation are already digital
  • routine code, tests and technical text can be generated automatically
  • many outputs can be checked with tests and static analysis

What protects the occupation

  • choosing the right objective, data and model-quality criteria
  • validation of assumptions and accountability for recommendations
  • a large share of value comes from human judgement, trust or personal accountability
  • architecture decisions and production accountability
  • understanding business constraints and legacy systems
  • security and complex system integration

United States · before tax

Average pay and outlook

Estimated nominal monthly pay. The scenario reflects sector, demand and task resilience; it is not a guarantee of income.

2026$11 950per month
2035$17 300scenario estimate
2026: $11 95020262027: $12 45020272028: $12 95020282029: $13 50020292030: $14 10020302031: $14 65020312032: $15 30020322033: $15 90020332034: $16 60020342035: $17 3002035
Find jobs ↗The partner link may contain referral parameters. Data benchmark: U.S. Bureau of Labor Statistics ↗

Job search and transitions

Find work with a lower risk index

Compare adjacent paths and open a current search for the selected occupation.

Skills you already have

  • data work
  • hypothesis testing
  • model-quality evaluation
  • valuation
  • return and risk analysis
  • systems thinking

Skills to build

  • AI-system evaluation
  • model-behavior monitoring
  • AI governance
  • financial modelling
  • AI-assisted scenario analysis
  • AI-agent-assisted development
CurrentAI Evaluation Engineer · 22%
→
Keepdata work · hypothesis testing · model-quality evaluation
→
AddAI-system evaluation · model-behavior monitoring · AI governance

Sources and update

Updated: 2026-09-25 · Assessment confidence: Medium

International Labour Organization · 2025Generative AI and Jobs: A Refined Global Index of Occupational Exposure

Nearly 30,000 tasks; a global assessment of GenAI exposure across detailed occupational groups.

World Economic Forum · 2025The Future of Jobs Report 2025

A survey of more than 1,000 employers representing over 14 million workers across 55 economies.

O*NET Resource Center · 2026O*NET Database

Occupational, task, skill and work-context structure.