6 templates across ML engineer, senior, MLOps, AI engineer, specialist, and founding hire, with the role disambiguation and compliance guidance the template farms skip. Download as DOCX.
Machine learning engineer is one of the blurriest titles in tech hiring. It overlaps with data scientist, AI engineer, data engineer, MLOps engineer, and domain specialists, and the fast-rising AI engineer role is quietly absorbing a lot of what people used to call ML engineering. So the first job of any machine learning engineer job description is to pin down which role you actually need, and the second is to get the compliance right, because AI work carries considerations the template farms ignore.
At FirstHR, we build hiring templates that name the parts the generic templates skip. For an ML role, that means clear role disambiguation plus the real compliance layer: FLSA classification, export controls, and the EU AI Act. The six below cover ML engineer and its main siblings, plus a startup founding-hire version. The guide to writing a job description covers the fundamentals.
TL;DR
A machine learning engineer builds, trains, and ships ML models into production, and the title overlaps with data scientist, AI engineer, and MLOps. There is no dedicated federal occupation code, so data scientists (median $112,590, May 2024) is the closest proxy; real ML pay runs far higher. The role is FLSA exempt under the computer employee exemption, and AI work raises export-control and EU AI Act questions.
What Is a Machine Learning Engineer?
A machine learning engineer is a software-engineering-centric specialist who designs, builds, trains, and deploys machine learning models into production systems. The role turns data and models into reliable, production-grade features, spanning the lifecycle from training through deployment and monitoring.
There is no dedicated federal occupation code for the role, so labor data captures it under proxies such as data scientists and software developers, which only approximate it. Most ML engineers come from computer science or engineering backgrounds and combine strong programming with applied machine learning. The title blurs with several adjacent roles, so a precise posting matters.
ML Engineer vs Data Scientist vs AI Engineer
The single most useful thing a machine learning job description can do is be clear about which role it means, because the titles genuinely overlap and attract different candidates.
Machine Learning Engineer
Builds and trains models
The core role: a software-engineering-centric specialist who designs, trains, and ships machine learning models into production. Heavier on math and model building than a data scientist.
Data Scientist
Insight and analysis
Focuses on analysis, experimentation, and insight from data, often with a broader background. Where the ML engineer productionizes models, the data scientist tends to explore and explain the data.
AI Engineer
Applied foundation models
The fastest-growing sibling: a software engineer integrating pre-trained foundation models with prompting, retrieval, and fine-tuning, rather than training models from scratch.
Data / MLOps Engineer
Pipelines and infrastructure
Data engineers build the data pipelines that feed models; MLOps engineers build the deployment and monitoring infrastructure that keeps models running reliably in production.
AI Engineer Is the Fastest-Growing Sibling
If your need is product features built on top of pre-trained foundation models, with prompting, retrieval, and fine-tuning, an AI engineer posting often fits better than a machine learning engineer one. AI engineer is the fastest-growing of these titles. Reserve machine learning engineer for roles that build and train models more directly, and use data scientist for analysis-and-insight roles.
Machine Learning Engineer Duties and Responsibilities
A machine learning engineer's duties cluster into modeling, production and pipelines, collaboration, and engineering quality. The balance shifts by role, more infrastructure for MLOps, more foundation-model work for an AI engineer, but these areas hold across the family.
Modeling
Design, build, train, and evaluate models
Optimize for accuracy, latency, and cost
Apply the right techniques to the problem
Production and pipelines
Deploy models to production and monitor them
Build data and feature pipelines
Manage versioning and reproducibility
Collaboration
Work with software, data, and product teams
Partner with leadership on ML direction
Mentor and review other engineers
Engineering quality
Write clean, tested, maintainable code
Document models, data, and experiments
Keep current with relevant tools and methods
A general ML engineer spans the whole lifecycle; an MLOps engineer focuses on deployment infrastructure; an AI engineer builds on foundation models. For a structured way to scope the role, the guide to defining job responsibilities walks through the process.
Which Template Should You Use?
Pick the template by role and seniority. The general ML engineer version is the flagship; the senior, MLOps, AI engineer, specialist, and startup versions match different needs and stages. Use this guide to choose.
Machine Learning Engineer
Core, end-to-end
The flagship: building, training, and deploying models across the lifecycle, with the FLSA and AI-compliance notes built in.
Senior / Staff ML Engineer
Leads and mentors
For a lead engineer owning architecture and production systems end to end, setting standards and mentoring the team.
MLOps Engineer
Pipelines and deployment
For the infrastructure and pipelines that get models into production reliably, sitting between ML and platform engineering.
AI Engineer
Applied foundation models
For building features on pre-trained models with prompting, retrieval, and fine-tuning, the fastest-growing sibling role.
NLP / Computer Vision
Domain specialist
For deep, domain-specific ML in natural language or vision, applying specialized techniques to text, speech, image, or video.
Startup / Founding
First technical hire
For an early-stage company hiring an early or founding ML engineer to own ML end to end, with the IP and compliance setup made explicit.
Match the Template to the Need
Build and train models end to end: Machine Learning Engineer. Lead and mentor: Senior / Staff. Pipelines and deployment: MLOps. Features on foundation models: AI Engineer. Deep NLP or vision work: NLP / Computer Vision. An early-stage company making a founding technical hire: Startup / Founding. Whichever you pick, classify the role as exempt and address the AI-specific compliance points.
Download all six as a single Word document or copy individual templates. Each follows the same structure: company and position summary, key responsibilities, qualifications, compliance notes, an EEO statement, and compensation. Fill in the brackets and post.
Download All 6 Templates
ML engineer, senior, MLOps, AI engineer, specialist, and startup. All in one DOCX.
Template 1: Machine Learning Engineer (General)
The flagship: building, training, and deploying models across the lifecycle, with the FLSA and AI-compliance notes built in.
This is the part the template farms skip entirely, and for an ML role it is where a careful posting adds real value. The compliance picture has a few moving parts worth getting right.
FLSA: almost always exempt
Machine learning engineers nearly always qualify as exempt under the FLSA computer employee exemption, which covers computer systems analysts, programmers, software engineers, and similarly skilled workers paid at least $684 per week on a salary basis or $27.63 an hour. ML salaries clear this easily, and the learned professional exemption can also apply.
Export controls and deemed exports
For controlled AI source code, model parameters, or chips, export-control rules can apply. Giving a foreign-national employee access to controlled technology can count as a deemed export requiring a license, and recent federal rules expanded controls on advanced AI models and chips. Flag this for defense-adjacent or frontier work.
IP assignment and confidentiality
Proprietary models, training data, and weights are core intellectual property, so IP assignment and confidentiality agreements are critical for ML hires. Non-compete enforceability varies by state and faces federal scrutiny, so rely on IP and confidentiality terms rather than broad non-competes.
EU AI Act exposure
The EU AI Act reaches beyond Europe: a US company whose AI outputs reach the EU can be a provider or deployer with transparency or high-risk obligations on a phased timeline. HR and employment AI is a named high-risk category, which matters if you build or use AI in hiring.
The computer employee exemption is detailed in DOL Fact Sheet 17E and the underlying regulation at 29 CFR 541.400. Export-control questions on AI technology run through federal rules administered by the Commerce Department.
Flag the AI-Specific Risks Early
For most small companies, the practical steps are simple: classify the ML engineer as salaried exempt, get IP assignment and confidentiality agreements signed before granting access to models and data, and flag export-control and EU AI Act questions where the work is frontier, defense-adjacent, or EU-facing. Get specific legal advice for those cases rather than relying on a template. This is general information, not legal advice.
Requirements and Qualifications
This is a skill-driven engineering role. Name the specific stack and experience your work requires, and keep the list focused, since long wish lists deter strong candidates.
Requirement
What to know
Education
Bachelor's in CS, engineering, or math typical; equivalent experience accepted
Core skills
Strong software engineering and Python; ML fundamentals
Frameworks
PyTorch, TensorFlow, scikit-learn, or similar per your stack
Infrastructure
Cloud and MLOps tooling for training, deployment, monitoring
Specialization
Foundation models for AI engineer; NLP or vision for specialists
Classification
Salaried, FLSA exempt under the computer employee exemption
Keep the must-have stack and experience clear, and tailor seniority and specialization to the role. The guide to writing a job description covers how to structure the rest.
Pay and Hiring Outlook
Machine learning engineers are among the highest-paid engineering roles, and federal proxy data understates their pay.
BLS Proxy Benchmark (May 2024)
Since there is no ML-engineer occupation code, the closest proxies are data scientists, median $112,590 a year with employment projected to grow 34% from 2024 to 2034 (one of the fastest-growing occupations), and software developers, median $133,080. National compensation surveys put actual ML engineer total compensation well above these proxies, often into the high six figures at well-funded firms, reflecting equity and a specialization premium (U.S. Bureau of Labor Statistics).
Budget against market data for the specific role and your stage rather than the federal proxy, and remember that equity is often a large part of total compensation, especially at startups. Market data shows wide variation driven by company stage, location, and specialization.
Hiring a Machine Learning Engineer for a Startup or Small Team
The honest picture: the titles blur so define the role, the compliance layer the templates skip is real, and the small-company version of this hire is a venture-backed startup making a founding technical hire. Here are the three realities to get right.
ML engineer, AI engineer, data scientist: the titles blur, so define the role
Machine learning engineer is not a cleanly standalone title; it overlaps heavily with several adjacent roles, and naming the wrong one attracts the wrong candidates. An ML engineer is software-engineering-centric and builds, trains, and ships models into production. A data scientist focuses more on analysis and insight from data. An AI engineer, the fastest-growing sibling, is increasingly a software engineer who integrates pre-trained foundation models with prompting, retrieval, and fine-tuning rather than training models from scratch, and for many product teams that is actually the role they need. A data engineer builds the pipelines that feed models, an MLOps engineer builds the deployment and monitoring infrastructure, and NLP or computer vision engineers are domain specialists. Before writing the posting, decide which of these you are actually hiring, because asking for a from-scratch model builder when you need someone to integrate an existing model wastes everyone's time. If your need is product features on top of foundation models, an AI engineer posting will likely fit better than a machine learning engineer one.
The compliance layer the templates skip: FLSA, export controls, EU AI Act
Generic templates list duties and stop, but ML roles carry compliance considerations worth getting right. On classification, an ML engineer is almost always exempt under the FLSA computer employee exemption, which applies to skilled computer workers paid at least the standard weekly salary or $27.63 an hour; ML salaries clear that threshold easily, so the role is overtime-exempt. Two less obvious issues matter for AI work specifically. First, export controls: controlled AI source code, model weights, and chips can be subject to export rules, and giving a foreign-national employee access to controlled technology can count as a deemed export requiring a license, which recent federal rules have tightened for advanced models and chips. Second, the EU AI Act reaches US companies whose AI outputs reach the EU, with transparency and high-risk obligations phasing in, and HR or employment AI is a named high-risk use. On top of these, IP assignment and confidentiality are critical because models and training data are core IP. None of the template farms address this, which is the one place a careful posting adds real value. This is general information, not legal advice.
The hiring company is usually well-funded; the small-team version is a startup
Be realistic about who hires an ML engineer. The role concentrates in big technology companies and well-funded AI startups, with compensation that runs well into six figures of total pay, and AI investment and hiring are hyper-concentrated in a few markets. A genuine small company hiring an ML engineer is typically a venture-backed, AI-native startup making an early or founding technical hire, a team of a handful of people moving fast, rather than a traditional Main Street small business. For that early-stage company, the first ML hire is often the moment it sets up formal HR and compliance for the first time, and the priorities are getting IP assignment and confidentiality right, classifying correctly, and onboarding cleanly so the engineer is productive fast. That is where a simple, flat-rate system helps a small team: e-signature for the offer and IP and confidentiality agreements, document management to store signed agreements and any compliance records, training modules for security and policy onboarding, task workflows so every hire runs the same way, and a simple HRIS and org chart as the team grows. Because pricing is flat rather than per seat, a small startup pays one rate as it scales. FirstHR does not run payroll or administer benefits, so pair it with a payroll provider. Applicant tracking is coming soon.
After You Hire: Onboarding a Machine Learning Engineer
Onboarding an ML engineer means getting the legal and security setup right first, then making the engineer productive fast. Send the offer stating the compensation, any equity, and the salaried exempt classification, collect the signed offer, and complete Form I-9 and tax forms as part of the new hire paperwork.
Because ML work centers on proprietary models and data, prioritize the IP assignment and confidentiality agreements and get them signed before the engineer accesses sensitive systems. Handle the compliance flags that apply: check export-control and deemed-export considerations for controlled AI technology, and note any EU AI Act obligations if your outputs reach the EU. Provision code, data, and cloud access to least privilege, and deliver security and policy training. Keep the signed onboarding documents and agreements on file. If this is an early hire at a young company, the guide to startup hiring covers the broader steps.
Because an early ML hire is often a small company's first formal hire, a documented, repeatable process saves real time. FirstHR fits the surrounding workflow directly: e-signature for the offer, IP, and confidentiality agreements, document management to store signed agreements and compliance records, training modules for security and policy onboarding, task workflows so every hire runs the same way, and a simple HRIS with an org chart as the team grows. Because pricing is flat rather than per seat, a small startup pays one rate as it scales. FirstHR does not run payroll or administer benefits, so pair it with a payroll provider. Applicant tracking is coming soon to FirstHR.
Key Takeaways
Machine learning engineer overlaps heavily with data scientist, AI engineer, data engineer, and MLOps; define which role you actually need before posting.
AI engineer is the fastest-growing sibling and often the right title when you need features built on pre-trained foundation models rather than models trained from scratch.
There is no dedicated federal occupation code; data scientists (median $112,590, May 2024) and software developers (median $133,080) are the closest proxies, and real ML pay runs higher.
ML engineers are almost always FLSA exempt under the computer employee exemption, given both their duties and their pay.
AI work raises compliance issues the templates skip: export-control deemed-export rules, the EU AI Act for EU-facing outputs, and IP assignment for proprietary models and data.
The small-company version of this hire is usually a venture-backed AI startup making a founding technical hire, not a traditional small business.
Frequently Asked Questions
What is a machine learning engineer?
A machine learning engineer is a software-engineering-centric specialist who designs, builds, trains, and deploys machine learning models into production systems. The role sits at the intersection of software engineering and applied machine learning: the engineer takes data and models and turns them into reliable, production-grade features that run at scale. There is no dedicated federal occupation code for machine learning engineer, so in labor data the role is captured under proxies such as data scientists, software developers, and computer and information research scientists, which means published figures only approximate the specific role. Most machine learning engineers come from computer science or engineering backgrounds and combine strong programming skills, particularly in Python, with a solid grasp of machine learning fundamentals, data handling, and the model lifecycle from training through deployment and monitoring. The title blurs with several adjacent roles, including data scientist, AI engineer, data engineer, MLOps engineer, and domain specialists in natural language processing or computer vision, so a precise job description matters. When you hire, the key decisions are which of these adjacent roles you actually need, what seniority the work requires, and whether you need someone who builds models from scratch or integrates pre-trained foundation models.
What does a machine learning engineer do?
A machine learning engineer builds machine learning systems and gets them running reliably in production. The work clusters into a few areas. On modeling, the engineer designs, builds, trains, and evaluates models, and optimizes them for accuracy, latency, and cost. On production and pipelines, the engineer deploys models to production, monitors their performance and drift, builds the data and feature pipelines that feed training and inference, and manages model versioning and reproducibility. On collaboration, the engineer works with software, data, and product teams, partners with leadership on machine learning direction, and at senior levels mentors and reviews other engineers. On engineering quality, the engineer writes clean, tested, maintainable code, documents models, data, and experiments, and stays current with relevant techniques and tooling in a fast-moving field. The exact mix depends on the role: a general ML engineer spans the whole lifecycle, an MLOps engineer focuses on the deployment and monitoring infrastructure, an AI engineer builds features on top of pre-trained foundation models, and an NLP or computer vision engineer applies deep domain-specific expertise. Across all of them, the role combines machine learning skill with the software engineering discipline needed to ship reliable systems, not just experimental notebooks.
What is the difference between a machine learning engineer and a data scientist?
The two roles overlap but differ in focus: a machine learning engineer is more software-engineering-centric and builds production systems, while a data scientist focuses more on analysis and insight from data. A data scientist typically explores data, runs experiments, builds analytical models, and communicates findings to inform decisions, often coming from a more varied academic background. A machine learning engineer takes models and turns them into reliable, deployed, production-grade systems, with a stronger emphasis on software engineering, infrastructure, and the full model lifecycle. In practice the line is blurry and varies by company, and small teams sometimes combine both in one role. The title also blurs with others. An AI engineer, the fastest-growing of these titles, is increasingly a software engineer who integrates pre-trained foundation models using prompting, retrieval, and fine-tuning, rather than training models from scratch, which is a meaningfully different job from a traditional ML engineer. A data engineer builds the data pipelines that feed models, and an MLOps engineer builds the deployment and monitoring infrastructure. When you write a posting, do not assume the titles are interchangeable. Decide whether you need analysis and insight, from-scratch model building, foundation-model integration, or data and deployment infrastructure, and name the role accordingly.
Is a machine learning engineer exempt or non-exempt from overtime?
A machine learning engineer is almost always exempt from overtime under the Fair Labor Standards Act, typically under the computer employee exemption. That exemption applies to computer systems analysts, computer programmers, software engineers, and other similarly skilled workers in the computer field whose primary duties involve systems analysis, design, development, and similar work, and who are paid either on a salary basis at the standard weekly threshold or on an hourly basis at a rate of at least $27.63 an hour. Machine learning engineer compensation clears the salary threshold many times over, so the pay test is easily met, and the duties of designing, building, and deploying machine learning systems fit squarely within the exemption. The role may also qualify under the learned professional exemption given the advanced knowledge typically required. As a practical matter, machine learning engineers are properly classified as salaried exempt employees and are not entitled to overtime. As always, job titles alone do not determine exempt status; the actual duties and compensation must meet the legal tests, and you should apply the higher of the federal or state requirements where a state sets stronger standards. This is general information, not legal advice.
Are there export-control or AI-regulation issues when hiring a machine learning engineer?
Yes, and they are easy to overlook. Two areas matter for AI work specifically. The first is export controls. Advanced AI source code, model parameters, and chips can be subject to US export-control rules, and a key concept is the deemed export: giving a foreign-national employee access to controlled technology inside the US can be treated as an export to that person's country and may require a license. Recent federal rules have expanded controls on advanced AI models and chips, so this is most relevant for frontier or defense-adjacent work, but it is worth checking before granting a new hire access to controlled technology. The second is the EU AI Act, which reaches beyond Europe. A US company whose AI systems or outputs reach the EU can be considered a provider or deployer with transparency or high-risk obligations on a phased timeline, and AI used in hiring and employment is a specifically named high-risk category, which is directly relevant if your machine learning engineer builds AI that touches HR decisions. On top of these, IP assignment and confidentiality agreements are essential because models and training data are core intellectual property. For most small companies the practical step is simply to flag these areas and get advice where the work is frontier, defense-adjacent, or EU-facing. This is general information, not legal advice.
What qualifications does a machine learning engineer need?
A machine learning engineer typically needs a strong software engineering foundation plus applied machine learning skills, usually anchored by a bachelor's degree in computer science, engineering, mathematics, or a related field, though equivalent practical experience is widely accepted and some roles prefer a master's or doctoral degree. The core technical qualifications are strong programming ability, especially in Python, solid command of machine learning fundamentals and data handling, and hands-on experience building and deploying models with common frameworks such as PyTorch, TensorFlow, or scikit-learn. Familiarity with cloud platforms and MLOps tooling for training, deployment, and monitoring is increasingly expected, since the job is about shipping production systems, not just prototyping. Beyond the baseline, requirements scale with the role and seniority: a senior or staff engineer needs a track record of leading projects and mentoring, an MLOps engineer needs deep infrastructure, CI/CD, and orchestration skills, an AI engineer needs experience with foundation models, prompting, retrieval, and fine-tuning, and a specialist needs domain depth in natural language processing or computer vision. Strong communication and collaboration matter across all of them, because the role works closely with software, data, and product teams. When you write the posting, separate the genuine must-haves, the specific stack and experience your work requires, from the long wish lists that deter good candidates.
Do small companies and startups hire machine learning engineers?
Yes, but the small companies hiring machine learning engineers are usually venture-backed, AI-native startups rather than traditional small businesses. Machine learning hiring is concentrated in large technology companies and well-funded AI startups, and AI investment is hyper-concentrated in a small number of companies and markets. When a genuinely small company hires an ML engineer, it is typically an early or seed-stage AI startup making a founding or early technical hire, often a team of just a handful of people, where that engineer owns machine learning end to end and has outsized impact. These companies move fast, run lean, and value versatility and ownership over deep specialization at the earliest stage. They are also the point at which a young company often sets up formal HR and compliance for the first time, since this may be one of its first full-time hires. So while machine learning engineer is not a role that traditional Main Street small businesses hire, small AI startups absolutely do, and for them the priorities around the hire are getting IP assignment and confidentiality right, classifying the role correctly, and onboarding cleanly and fast. The startup and founding-hire template on this page is written specifically for that early-stage, small-team reality.
What happens after I hire a machine learning engineer?
Run a structured onboarding that gets the legal and security setup right up front, then makes the engineer productive quickly. Start with the employment basics: send the offer stating the compensation, equity if any, and the salaried exempt classification, collect the signed offer, complete Form I-9 in the first days, and gather the W-4 and any state tax forms. Because machine learning work centers on proprietary models and data, prioritize the intellectual property and confidentiality agreements and get them signed before the engineer accesses sensitive systems, since models, weights, and training data are core IP. Handle the compliance flags that apply to your work: if you deal in controlled AI technology, check export-control and deemed-export considerations before granting a foreign-national hire access; if your AI outputs reach the EU, note any EU AI Act obligations. Then handle security and access onboarding, including code, data, and cloud access provisioned to the least privilege needed, and deliver security and policy training. Finally, orient the engineer to your codebase, data, model stack, and the software, data, and product people they will work with, and set clear early goals. Because an early ML hire is often a small company's first formal hire, having this documented and repeatable saves real time. FirstHR fits the surrounding workflow directly: e-signature for the offer, IP, and confidentiality agreements, document management to store signed agreements and compliance records, training modules for security and policy onboarding, task workflows so every hire runs the same way, and a simple HRIS with an org chart as the team grows. FirstHR does not run payroll or administer benefits, so pair it with a payroll provider. Applicant tracking is coming soon to FirstHR.
