Hire machine learning engineers
A machine learning engineer builds, trains, evaluates and deploys machine learning models, owning the data pipelines, training runs and serving infrastructure that turn a model into a dependable part of a product.
Checked
Machine learning engineers sit on the model side of what Latent Space, the newsletter that named the AI engineer role in June 2023, calls the API line: they train, host and evaluate models, while AI engineers build products on models reached through an API. Latent Space calls the line permeable, and many job posts ask for some of both.
The title pays more than most engineering titles. Levels.fyi puts the US median total compensation for machine learning engineers at $282,950 across 1,065 self-reported salaries, updated on 31 October 2026, against $199,190 for software engineers overall. A wrong hire at that price is an expensive one.
What a machine learning engineer does
Builds the data pipeline
Collects, cleans and labels training data, and keeps training data and production inputs consistent so the model sees in production what it saw in training.
Trains and fine-tunes models
Chooses the model family, trains or fine-tunes it, and tracks every run so a result can be reproduced and compared.
Evaluates against the business problem
Builds held-out test sets and metrics that reflect what the product needs, not only a public benchmark score.
Serves and monitors the model
Deploys the model behind an API or on a device, watches latency, cost and drift, and retrains when the inputs change.
Machine learning engineer vs AI engineer
The two roles share tools and often share a team, but they start from different places. A machine learning engineer starts from data and produces a model. An AI engineer starts from a model someone else trained, usually a large language model behind an API, and produces a product feature: prompts, retrieval, tool calls and the evals around them.
Pay data shows the split as well. On Levels.fyi, the Machine Learning Engineer title reports a US median of $282,950 (1,065 salaries, updated 31 October 2026) and the AI Engineer title $156,000 (339 salaries, updated 1 November 2026). The figures are self-reported and the mix of employers behind each title differs, so the gap is a signal rather than a price list.
A company that needs a computer-vision model, a forecasting model or a recommender needs machine learning engineering. A company adding an assistant, an agent or document extraction on top of a foundation model usually needs an AI engineer first, and a machine learning engineer only once it fine-tunes or hosts its own models.
Ways to hire a machine learning engineer
The routes differ less in headline cost than in who keeps the knowledge afterwards. The figures below were checked on 1 November 2026.
Full-time hire
Levels.fyi reports a US median total compensation of $282,950 for the title (p25 $200,000, p75 $385,500, p90 $495,000), counting stock and bonus. Recruiting From Scratch read 999 ML engineer job posts in June 2026 and found a median posted salary of $220,000 (p25 $185,000, p75 $257,000).
Recruiter
Paraform charges 25% of first-year base on hire, with a 90-day replacement guarantee. On a $220,000 base that fee is $55,000, and the search still ends in a full-time hire with the same timeline.
Contractor or talent marketplace
Arc lists freelance rates from $15 to over $110 an hour, with a trial period of up to two weeks, and charges 20% of annual salary for a full-time placement. A short contract suits a scoped model build; the monitoring and retraining that follow need someone who stays.
Embedded lab
An independent team such as Motiontic builds the model and the product around it inside the client's codebase, then hands over the pipelines, evals and documentation. It costs more per month than a single contractor and takes on a limited number of projects at once.
Where machine learning engineers find jobs
AI company boards
Inference Jobs listed 8,366 jobs across 133 AI companies on 1 November 2026, with 343 in its Machine Learning Engineer category. aijobs.ai showed 555 openings under Machine Learning.
Startup boards
Wellfound showed 781 Machine Learning Engineer results, many with salary and equity ranges on the listing. Y Combinator's job board and the monthly Hacker News Who is hiring thread carry startup roles posted by founders.
Communities
The MLOps Community runs a Slack, meetups and a job board that lists more than 250 open AI roles.
General boards
LinkedIn carries the largest volume, and Built In shows the posted salary range on most listings.
Working with Motiontic
Both sides of the API line
Motiontic's engineers train and serve a model where the problem needs one, and build on a foundation model where that is cheaper and accurate enough. The choice is made per problem, not per job title.
Decided on the client's data
Whether to train, fine-tune or call a hosted model is settled by testing on the client's own data for accuracy, latency and cost. Motiontic is not tied to a model provider or a cloud.
Production code in week one
Engineers join the client's codebase and team chat once the scope is agreed, and the first production code lands in the first week.
A handover the team can run
Data pipelines, evals, monitoring and documentation stay with the client's engineers, followed by continued support.
Field note
Computer-vision quality control on a steel line
Motiontic's engineers built computer-vision measuring and quality control on the assembly line at a steel plant. It replaced manual vernier-caliper checks on steel plates and made inspection autonomous.
+40% productivity
Ways to hire a machine learning engineer (checked 1 November 2026)
| Route | Cost | Time to first work | Trade-off |
|---|---|---|---|
| Full-time hire | Levels.fyi US median total pay $282,950; median posted salary $220,000 (Recruiting From Scratch) | A search of weeks to months, then onboarding | Knowledge stays in-house; one hire rarely covers data, training, serving and product |
| Recruiter | Paraform: 25% of first-year base, paid on hire | The same search, with reach into passive candidates | Ends in the same full-time hire, with the fee on top |
| Contractor marketplace | Arc: $15 to $110+ an hour, or 20% of annual salary for a placement | Once a match is found | Suits a scoped build; monitoring and retraining need someone who stays |
| Embedded lab (Motiontic) | Monthly engagement, scoped per project | Production code in the first week after scope | Vendor-neutral team with handoff; more per month than one contractor |
Common questions
What does a machine learning engineer do?
A machine learning engineer builds the data pipelines, trains and evaluates models, and deploys and monitors them in production, retraining when the inputs change.
How much does a machine learning engineer earn?
Levels.fyi reports a US median total compensation of $282,950 for machine learning engineers, from 1,065 self-reported salaries updated on 31 October 2026. Job posts list less, because they show salary only: Recruiting From Scratch found a median posted salary of $220,000 across 999 posts in June 2026.
What is the difference between a machine learning engineer and an AI engineer?
A machine learning engineer trains, evaluates and serves models. An AI engineer builds product features on models trained by someone else, usually through an API, with prompts, retrieval, tool calls and evals.
Does a company need a machine learning engineer to build with large language models?
Often not at first. A product built on a hosted foundation model mainly needs AI engineering. Machine learning engineering becomes necessary when the company fine-tunes or trains its own models, or serves them on its own infrastructure.
How much does it cost to hire a machine learning engineer through a recruiter?
Paraform charges 25% of first-year base on hire, and Arc charges 20% of annual salary for a full-time placement. Both fees come on top of the salary.
Related
AI Engineer vs ML Engineer
Building product on hosted models, compared with training and serving the models.
AI Engineer
Builds the model-facing parts of a product: prompts, retrieval, tools, evals and the code around them.
Applied AI Engineer
Turns frontier models into product features, and measures whether they are good enough to ship.
RAG vs Fine-Tuning
Fetching the right facts at query time, compared with retraining the model's behaviour.
AI product development
New products and major AI features, built from scratch alongside the client's team.
Sources
- Levels.fyi: Machine Learning Engineer pay (US, 2026-10-31)
- Levels.fyi: AI Engineer pay (US, 2026-11-01)
- Levels.fyi: Software Engineer pay (US, 2026-10-31)
- Recruiting From Scratch: ML engineer salary in 2026 (2026-06-25)
- Paraform: How to hire a forward deployed engineer (2026-09-04)
- Arc pricing
- Latent Space: The Rise of the AI Engineer (2023-06-30)
- Inference Jobs
- aijobs.ai
- Wellfound: Machine Learning Engineer jobs
- MLOps Community
Running agents that have to hold up?
Describe the agent, the systems it touches and what it must never get wrong. The first call is with a lab engineer, who will say plainly whether it is ready for production, and whether Motiontic is the right team.