Machine Learning Engineer Salary: Federal Pay Data, Top States, and What an Online AI Degree Costs

Machine learning engineer salary figures come from the closest federal group, data scientists, because BLS has no line for the title itself. Their May 2025 median wage was $120,230 and the mean was $126,800. Pay runs from $67,240 at the 10th percentile to $199,130 at the 90th, and your state can move it by tens of thousands of dollars.

What Does a Machine Learning Engineer Earn?

BLS counts machine learning engineers with data scientists, so these figures cover that wider group rather than machine learning engineers alone. It is the occupation whose duties overlap most: creating, validating, testing and updating models, then explaining the results.

Here are the national May 2025 wages.

• 10th percentile: $67,240.

• 25th percentile: $85,660.

• Median: $120,230, or $57.80 an hour.

• Mean: $126,800, or $60.96 an hour.

• 75th percentile: $158,880.

• 90th percentile: $199,130.

The middle half of the group earns between $85,660 and $158,880. That is a gap of $73,220 inside one occupation, and it is the most useful fact to hold in mind when you compare offers.

For the day-to-day work behind these numbers, see our page on machine learning engineer jobs.

How Do Percentiles Show Experience?

The federal wage survey does not break pay out by years of experience, and we will not invent bands. Percentiles are the honest stand-in, with a catch: they describe everyone in the group at once.

Think of it this way. The 10th percentile is what the lowest-paid tenth earn or less, and the 90th is what the highest-paid tenth earn or more. People early in a career tend to cluster lower in that range, and people with deep experience or scarce skills tend toward the top.

So use the range like this. If you are changing careers, plan around the 25th percentile, $85,660, until a real offer says otherwise. If you already work in software or analytics, the median, $120,230, is a fair first benchmark. The 75th percentile, $158,880, is where strong offers in competitive states start to show up.

Mean is higher than median because a few very high wages pull the average up. When a job post quotes a single number, ask which one it is.

Where Do Machine Learning Engineers Earn the Most?

State is the biggest lever you can see. Here are the states with the highest annual mean wages for the group, with their medians and job counts.

• Washington: mean $158,040, median $163,350, 9,600 jobs.

• California: mean $156,000, median $141,590, 39,310 jobs.

• Maryland: mean $144,080, median $136,370, 3,340 jobs.

• District of Columbia: mean $142,580, median $126,490, 2,680 jobs.

• New Jersey: mean $138,590, median $135,280, 6,430 jobs.

• New York: mean $137,690, median $130,460, 23,970 jobs.

• Massachusetts: mean $136,460, median $131,750, 9,420 jobs.

• Virginia: mean $134,590, median $126,430, 8,920 jobs.

Washington also has the highest median, $163,350. At the low end, Mississippi has a median of $69,490, North Dakota $81,900 and Louisiana $78,760.

Cost of living matters before you chase the top. Rent and taxes differ by state, and a higher wage in an expensive metro can feel smaller than a lower wage elsewhere. If you can work remotely for an employer in a high-pay state, ask whether your pay follows your location.

Where Are the Jobs?

Pay only helps if openings exist near you, or in a role you can do from home. BLS counted 262,440 jobs in the group nationally.

California has the most with 39,310. Texas follows with 25,860, New York with 23,970, Pennsylvania with 13,810 and North Carolina with 11,430.

Notice how the job list differs from the pay list. Texas has a median of $122,090, near the national figure, and Pennsylvania’s is $106,850. North Carolina’s is $119,090. Washington pays the most and has fewer jobs than any of them. California is the rare state that does both.

If you want the most openings, look at the job-heavy states. If you want the highest pay, look at the pay-heavy ones, and be prepared to compete for fewer seats.

What Else Moves Pay?

Some things we can show from the data, and some we cannot, so here is the split.

We can show the spread. Inside California, the 10th percentile is $77,480 and the 90th is $224,920. Inside Texas, they are $64,540 and $170,780. A top state raises the ceiling and the floor, but a wide gap remains.

We can show that the title is a family, not a single job. Employers may file model-building work under data scientist, software developer or research scientist. Pay differs by group, which is why the same skills can come with different offers. Our AI engineer salary page shows the other group.

We cannot show pay by employer type or degree. The federal data we use does not break it out, so we do not guess. If an employer type matters to you, check postings for it.

Does Pay Differ by Degree?

BLS does not report pay by degree for this group. What we can say is what the degrees ask of you and what they cost.

BLS lists a bachelor’s degree as the typical entry education for the data scientist group, and says graduate school is not required. Online master’s programs teach machine learning in more depth, and many working engineers choose one to move into model work.

A degree is not a salary ticket. It is one input, next to experience, projects and where you live. Price it like an investment with an uncertain return, and keep the cost low enough that a middling outcome is still fine.

What Does an Online AI Degree Cost?

Prices vary more than pay does. These are totals from our program data for online AI master’s programs, tuition only unless noted.

• The University of Texas at Austin: $9,990.

• University of Colorado Boulder: $15,750, with performance-based admission.

• Florida Atlantic University: $24,000.

• West Virginia University: $24,150 for 30 credits at $805 each, before fees.

• Saint Louis University: $24,300.

• Boston University: $25,000 for several master’s options.

• Purdue University: $28,500 for 30 credits at $950 each.

• University of Illinois Chicago: $32,256 for 36 credit hours.

• George Washington University: $37,500.

• Drexel University: $63,000 for 45 quarter credits at $1,400 each, before fees.

• Columbia University: $81,000.

To compare prices on one page, use our cheapest online AI degrees list. For master’s programs only, see the online master’s in AI ranking, and for programs that teach machine learning, the machine learning degrees ranking.

Capitol Technology University’s online PhD in Machine Learning costs $58,800 and is built for working professionals. Bachelor’s options include Dakota State University at $44,856 total.

How Does Degree Cost Compare With Pay?

Put a price next to a wage and the question gets concrete. The University of Texas at Austin’s total of $9,990 is less than a year at the 10th percentile, $67,240. Drexel’s $63,000 is close to that whole year of pay. Columbia’s $81,000 is more than the annual wage at the 10th percentile, and below the median wage of $120,230.

None of that is a payback period, because you do not know your own wage yet. But it tells you how much risk each price carries.

A lower price gives you more room. If you reach the median, even a modest degree pays for itself in a short time. If you land at the 25th percentile, a program like Texas at $9,990 or Purdue at $28,500 is easier to carry than Columbia at $81,000.

Two more costs sit outside the sticker. Many schools list per-credit tuition before fees, so the bill can be higher. And if you do not have the math or programming a program expects, you may need extra courses first. West Virginia University, Purdue and Drexel all list prerequisites, and Colorado Boulder uses performance-based admission with no bachelor’s required.

For programs that do not ask for the GRE, see the no-GRE master’s list.

What Do Wage Levels Look Like by the Hour?

Hourly figures help if you are comparing a contract role or a part-time schedule against a salary. The group’s median is $57.80 an hour and its mean is $60.96. In California the median is $68.07 an hour, in Texas $58.70 and in Pennsylvania $51.37.

Those are straight conversions of annual wages, so they do not include bonuses, equity or benefits. Federal wage data counts pay, not total compensation. If an offer includes stock or a bonus, ask for the dollar amounts in writing before you compare.

Why Do Offers for the Same Title Differ So Much?

Two postings with the same title can sit at opposite ends of the range, because the title says little about the work. One employer may want someone to maintain a single model in production, while another wants a person who designs new ones. Pay follows scope, location and what the team can afford.

When an offer comes in, ask what the first year of work looks like, who you report to and whether the role is on a research team, a product team or a data team. Then place the base pay against your state’s percentiles above.

How Fast Is the Field Growing?

BLS projects data scientist employment to grow 35 percent from 2025 to 2035, which it calls much faster than average. That is about 95,400 more jobs on a base of 275,600, with about 24,800 openings a year.

Growth is direction, not a guarantee. Projections are estimates built on assumptions, and BLS revises them. They also do not say who gets hired, and a posting can ask for a degree or experience you do not yet have.

Still, a growing group is easier to enter than a shrinking one, and that is the best reason to look at it.

What Does a Salary Number Not Tell You?

One figure hides a lot, so keep these in mind.

It does not tell you how long it takes to reach a given level. BLS reports medians for everyone in the group, not for new hires or for people two years in.

It does not tell you about the work. A machine learning role can mean tuning models, building pipelines, cleaning data or presenting results. Employers weight those differently.

It does not tell you about stability. A single company can hire or lay off regardless of national growth.

It does not tell you what your degree adds. We do not guess that either.

What it does tell you is the market’s range. If offers you see sit at the low end of it, you can ask why. If they sit well above, you can ask what the employer expects in return.

How Do You Use These Numbers?

Plan around the range, not the midpoint. If you are early in your career, plan for the lower half. If you have experience, plan for the upper half.

Check your state against the table above, then read three real postings in your area for the title you want and see which federal group they resemble.

Last, price your education against that range. Our guide on how to become an AI engineer lays out the degree paths, and the machine learning engineer jobs page covers what the work involves and how to get hired.