Automation index
Can AI do a ML engineer’s job?
Takes machine-learning models from a trial to a live system.
It also does not predict how many people will lose their jobs.
IS WORK OVER?
ML engineer
41%
AI or a robot can already do 41% of this job.
#10 / 182
isworkover.com
Where the week goes
Score = Σ (weight × saving). Weights are our frequency estimate, not an O*NET time %. Saving is separate evidence or estimate.
Task weights are our estimate from O*NET frequency (FT). O*NET does not publish share of the week.
| Task | Share | AI | Adds |
|---|---|---|---|
Train and evaluate
EstimateNo public source pins this saving. It is an estimate from the method. |
35% | 55% | +19.3 |
Shipping to prod
EstimateNo public source pins this saving. It is an estimate from the method. |
25% | 45% | +11.3 |
Data and features
EstimateNo public source pins this saving. It is an estimate from the method. |
25% | 30% | +7.5 |
The business problem
EstimateNo public source pins this saving. It is an estimate from the method. |
15% | 20% | +3 |
| Total | 100% | 41% |
Modelled week 100% · Public evidence 0% · Estimate 100% · Confidence: Low
Occupation sources · O*NET 15-2051.00 · ESCO/ISCO 2521