AI Engineer Salary US 2025: BLS Data

AI Engineer Salary in the US: What the Data Supports

Dan Patiño

· AI Strategy & Innovation at Coderhouse

Salary pages for this role tend to quote a single average and stop. The useful version needs three things: which occupation code the number comes from, what the bottom of the range looks like, and how much geography moves it. BLS publishes no occupation called 'AI Engineer'. The figures here use Software Developers (SOC 15-1252), with Computer and Information Research Scientists (SOC 15-1221) as a second reference point, from the OEWS May 2025 release.

National median

$135,980

Projected growth

15.8% through 2034

What is the median ai engineer salary?

The national median is $135,980 a year as of May 2025, against $140,300 for Computer and Information Research Scientists. BLS projects growth of 15.8% through 2034. Half of workers earn less than the median, so it describes a midpoint rather than a target.

What does the full range look like?

The spread matters more than the midpoint. Nationally the 10th percentile is well below the median. Across the 25 largest metros the 10th percentile ranges from roughly $63,000 to $110,000 depending on the market, a wider gap than most people get from changing jobs.

Where does the pay actually differ?

Highest-paying states by median: California $174,410; Washington $166,540; New York $166,180; Massachusetts $165,210; Oregon $142,720; New Hampshire $139,720. Lowest: Mississippi $95,330; South Dakota $96,310; Louisiana $102,800; Arkansas $104,490; North Dakota $105,600; Indiana $105,920. Among metro areas with at least 1,000 people in the occupation, the top payers are: San Jose-Sunnyvale-Santa Clara $213,110; San Francisco-Oakland-Fremont $186,640; Seattle-Tacoma-Bellevue $167,280; New York-Newark-Jersey City $166,830; Boston-Cambridge-Newton $166,090; Boulder $164,560. Cost of living is not adjusted in any of these figures, so a high nominal median in an expensive metro can be worth less in practice than a moderate one elsewhere.

What drives the differences?

Four things move pay in this field, roughly in order of impact early on: the metro you work in, the industry (finance and technology pay above healthcare and government for the same title), the size of the employer, and the scope of the role. Years of experience matters, but it compounds later rather than immediately. Note also that OEWS covers base wages only, so equity and bonuses, which are substantial at large technology employers, are not in any of these numbers.

Primary BLS occupation

15-1252 (Software Developers)

Secondary BLS occupation

15-1221 (Computer and Information Research Scientists)

Secondary occupation median

$140,300

Highest-paying states

California $174,410; Washington $166,540; New York $166,180; Massachusetts $165,210; Oregon $142,720; New Hampshire $139,720

Lowest-paying states

Mississippi $95,330; South Dakota $96,310; Louisiana $102,800; Arkansas $104,490; North Dakota $105,600; Indiana $105,920

Top-paying metros

San Jose-Sunnyvale-Santa Clara $213,110; San Francisco-Oakland-Fremont $186,640; Seattle-Tacoma-Bellevue $167,280; New York-Newark-Jersey City $166,830; Boston-Cambridge-Newton $166,090; Boulder $164,560

How do you move up the range?

The reliable levers are scope and evidence. Moving from producing reports to designing measurement, running experiments and defending a recommendation is what separates the middle of the range from the top of it. Coderhouse teaches that progression directly: the AI Engineering course is live online, 8 weeks long, $595 at launch price, with a curriculum reviewed by Google, Microsoft and MongoDB. It is a course, not a placement program, so there is no job guarantee attached.

AI Engineering

8 weeks · live online

Build AI systems you can explain, evaluate, and improve—not just demos that happen to call an LLM. Work hands-on with retrieval, RAG, tool use, agents, evals, observability, and deployment while learning the engineering judgment behind reliable AI applications.

See Dates

$595

Data Analytics

11 weeks · live online

Learn to turn data into decisions—not just dashboards. Build practical skills in SQL, Python, metrics, statistics, experimentation, visualization, and AI-assisted analysis, with live feedback from practitioners who do this work professionally.

See dates

$595

Frequently asked questions

How much does a ai engineer make in the US?

The closest BLS occupation shows a national median of $135,980 a year as of May 2025. BLS publishes no occupation called 'AI Engineer'. The figures here use Software Developers (SOC 15-1252), with Computer and Information Research Scientists (SOC 15-1221) as a second reference point, from the OEWS May 2025 release.

What is a realistic starting salary?

The 10th percentile is the best available proxy, since BLS does not publish by experience level. It sits well below the median, and it varies by metro considerably more than by employer.

Does location matter more than experience?

Early on, usually yes. The gap between the highest and lowest paying states exceeds what most people gain in their first several years in the role.

Do these figures include bonuses and equity?

No. OEWS covers base wages, which means total compensation at large technology employers is materially higher than these numbers suggest. Treat them as a floor for that segment.

How current is this data?

It is the May 2025 OEWS release, the most recent available. BLS updates it annually, and estimates draw on three years of collected panels, so they lag rapid changes in pay.

Methodology and caveats

BLS publishes no occupation called 'AI Engineer'. The figures here use Software Developers (SOC 15-1252), with Computer and Information Research Scientists (SOC 15-1221) as a second reference point, from the OEWS May 2025 release. OEWS estimates are produced from six semiannual panels collected over three years, which makes them stable but slower to reflect sudden market shifts. They measure wages where the job is located, cover base pay only, and exclude equity, bonuses and self-employed workers. Job titles in this field are not standardized, so any single occupation code is an approximation rather than an exact match for a title you see in a posting.

U.S. Bureau of Labor Statistics, OEWS May 2025