Skip to content
EmergedCore profileemerged c.2015

Machine Learning Engineer

Also: AI engineer

Builds and deploys machine learning systems.

ISCO-08 2512

1926

A · The occupation 100 years ago

Did not exist.

Historical imagery: not yet included. Images are added only after licensing (public domain or explicit permission) is confirmed.

1926–2026

B · The evolution

An occupation-specific milestone narrative has not yet been researched for this profile. The technology record below shows the documented developments that reshaped it.

What disappeared

  • No major tasks fully disappeared.

What technology took on

  • • Model development (from c.2023)
  • • Evaluation (from c.2024)

What remained fundamentally human

  • • Safety and deployment decisions

2026

C · The occupation today

High demand; building on foundation models; evaluation and safety work growing.

Required skills

  • ML
  • Software engineering

Typical tools & software

  • PyTorch
  • Cloud GPUs

Demonstrated today

Technology already used for these tasks (documented in 2015–2026)

  • • Model development
  • • Evaluation

Speculative capability

Not yet demonstrated at scale — these appear only in scenarios

  • • AI alignment oversight (central c.2030)

Barriers to automation (editorial assessment, 0–1)

These factors delay or cap automation in the model.

  • Regulation & licensing0.30
  • Physical environment0.05
  • Human accountability0.60
  • Social acceptance0.20
  • Economic viability0.75
  • Underlying demand trend+0.60

2030→2126

D · The future

Scenarios, not predictions. Confidence falls with the horizon.

2050 · CENTRAL SCENARIO

AI builds much AI; engineers focus on objectives, evaluation and safety.

Low confidence

2126 · CENTRAL SCENARIO

Unknown — possibly the most transformed occupation of all.

Speculative
0255075100192619501975200020262050207521002126NOWscenarios →← reconstructed
ConservativeCentralAcceleratedTask transformation index (0–100) · not a probability of job loss
Structured forecast per horizon and scenario (model task-curve-v1). Employment outlook is a qualitative direction, not a jobs number.
YearScenarioTransformationEmployment outlookSkills outlookConfidence
2030Conservative24Growing
Central20GrowingCore human skills remain in safety and deployment decisions; new capabilities needed for ai alignment oversight; plus fluency directing and auditing AI tools.Low
Accelerated21Growing
2050Conservative20Growing
Central25GrowingCore human skills remain in safety and deployment decisions; new capabilities needed for ai alignment oversight; plus fluency directing and auditing autonomous systems.Low
Accelerated40Growing
2075Conservative26Growing
Central43GrowingCore human skills remain in safety and deployment decisions; new capabilities needed for ai alignment oversight; plus fluency directing and auditing autonomous systems.Speculative
Accelerated52Stable, transforming
2126Conservative41Growing
Central52Stable, transformingCore human skills remain in safety and deployment decisions; new capabilities needed for ai alignment oversight; plus fluency directing and auditing autonomous systems.Speculative
Accelerated57Stable, transforming

Salary implications: not modelled. No credible occupation-level wage forecasts exist for these horizons, so we do not publish one.

Tasks

E · Occupational transformation

  • Non-routine analytic
  • Creative production
  • Judgement & accountability
2050
25/100Model output · Central scenarioLow confidence
  • Human-led · 1
  • Assisted · 2
  • Automated · 0
  • New task · 1
  • Retired · 0
Tasks in 2050
TaskMachine involvement
Model developmentAssisted
EvaluationAssisted
Safety and deployment decisionsHuman-led
AI alignment oversightnewHuman-led

How this is derived: tasks are classified with the routine/non-routine framework (Autor, Levy & Murnane 2003); each task type follows a capability curve per scenario, delayed by this occupation's adoption barriers (regulation, physical environment, social acceptance, economics) and capped by accountability. The index is a share of task content, not the probability that the job disappears. Full methodology →

Evidence

1 sources cited

AI Index Report 2025Stanford Institute for Human-Centered Artificial Intelligence · 2025-04Annual compilation of data on AI research, technical performance, investment, adoption and policy.Institutional recordID ai-index-2025Original ↗Register

Profile last updated 2026-11-17. This is a core profile: the occupation-specific historical narrative is still to be researched, so section B relies on documented technology records.