Compensation is where most deep-tech hiring processes quietly fail. A brief goes out priced against last year's market, the shortlist declines at offer stage, and the role sits open for another quarter.
This guide sets out the published U.S. wage benchmarks for the engineering occupations that AI, semiconductor, and energy employers hire into permanently — and explains how to read them before you set a band. Every figure below comes from the U.S. Bureau of Labor Statistics; sources are linked at the end.
U.S. median pay by engineering occupation
BLS reports a national median annual wage per occupation. Median means half of workers in that occupation earn more; it is not a starting salary and not a ceiling. Growth is the projected change in employment from 2024 to 2034.
| Occupation | Median annual wage | Projected growth, 2024–34 |
|---|---|---|
| Computer hardware engineers | $155,020 | +7% |
| Software developers, QA analysts, and testers | $131,450 | +15% |
| Electrical and electronics engineers | $118,780 | +7% |
| Data scientists | $112,590 | +34% |
| Materials engineers | $108,310 | +6% |
| Mechanical engineers | $102,320 | +9% |
| Industrial engineers | $101,140 | +11% |
| All architecture and engineering occupations | $97,310 | — |
| All U.S. occupations | $49,500 | — |
How these map to the roles you are actually hiring
Federal occupation codes are broader than job titles. A few translations that matter when you are pricing a brief:
- Chip design, verification, DFT, and physical design roles fall largely under computer hardware engineers ($155,020 median) — the highest-paid engineering occupation in the set.
- ML engineers, MLOps, and applied research engineers usually sit between software developers ($131,450) and data scientists ($112,590), depending on how much of the role is production systems versus modelling.
- Power systems, grid, substation, and BESS engineers map to electrical and electronics engineers ($118,780).
- Fab and process engineering roles split across materials engineers ($108,310), industrial engineers ($101,140), and mechanical engineers ($102,320).
Three adjustments to make before you set a band
A national median is a floor for planning, not an offer. Three factors move it materially:
- Location. BLS publishes state and metro wage estimates alongside each national figure, and the spread between a Bay Area or Austin band and a national median is significant. Price against the metro you are hiring into, not the national number.
- Seniority. The published percentile tables are more useful than the median for senior hires — the 75th and 90th percentile figures in the OEWS data are the realistic reference points for staff and principal-level engineers.
- Scarcity within the occupation. Data scientists are projected to grow 34% over the decade against 7% for hardware engineers. Fast-growing occupations clear above the median more often, because demand outruns supply faster than the wage data updates.
Why offers fail even at the right number
In our experience running permanent searches in these sectors, most lost offers are not lost on base salary alone. They are lost on total package clarity, equity explanation, relocation, and time. A candidate holding two offers takes the one they understand.
The practical fix is to fix the band before the search starts, not after the shortlist arrives — and to be able to state base, bonus, equity, and review cycle in one sentence at first conversation.
What to remember
Median U.S. pay for deep-tech engineering occupations runs roughly 2–3x the $49,500 national median across all occupations.
Computer hardware engineering is the highest-paid occupation in the set at a $155,020 median; data science is the fastest growing at +34% through 2034.
Adjust the national median for metro, seniority percentile, and occupational growth before you set a band.
Fix the salary band before the search starts — most lost offers are lost on package clarity and time, not base pay.