tech
Software Engineer: modeled AI cost and task mix
Designs, builds, and maintains software systems. Combines coding with system design, code review, and operational responsibility.
This page uses a representative geography and experience level. Build your own Wagecard to use your actual task mix, hours, location, experience, and optional salary.
- Operational AI cost
- Market pay (p50)
- Four task classes
- Qualitative confidence
- Documented methodology
Operational AI cost
Modeled AI running cost includes tokens, human review, retries, error cost, integration, and orchestration.
Market pay (US Tier 2, mid-career)
Market pay percentiles: p25 $112,299, p75 $159,654, p90 $189,420.
Economic substitution exposure
Lower scores mean a person is more reliable or cost-effective for more of the task time. Human-critical work.
Task mix
Four task classes
Hours-weighted share of this role's tasks
Task details
How each task was classified
For each task, see AI capability, reliability, error cost, and where a person remains more reliable or cost-effective.
| Task | Capability | Reliability | Error cost | Human-advantage |
|---|---|---|---|---|
| Debug production incidents | 50 | 45 | 5/5 | 60/100 |
| On-call response | 30 | 30 | 5/5 | 80/100 |
| Code review | 55 | 55 | 3/5 | 50/100 |
| Write production code | 78 | 70 | 3/5 | 25/100 |
| Write documentation | 80 | 75 | 2/5 | 25/100 |
| Write tests | 82 | 75 | 2/5 | 20/100 |
| Design system architecture | 45 | 40 | 5/5 | 75/100 |
| Mentor junior engineers | 20 | 20 | 5/5 | 90/100 |
Representative US Tier 2, mid-career example using capability matrix v1 and model v1-mvp. See the documented methodology at /methodology.
Build a Wagecard for your Software Engineer task mix
Use your actual tasks, hours, location, experience, and optional salary. Your Wagecard shows modeled AI running cost, market pay, the four task classes, and qualitative confidence on the core estimates.