Women in AI by FemTechConf
AI Careers in the USA: Roles, Skills and Hiring Markets
Choose an AI career path with clear evidence, understand regional differences and build a portfolio that can withstand the expectations of US employers.
+33.5%: Projected US data-scientist employment growth from 2024 to 2034, according to BLS. This is a projection for one occupation, not all AI jobs.
How strong is the US AI hiring opportunity?
US government labour projections indicate growth in several occupations relevant to artificial intelligence, but it is misleading to combine every one into a single measure of AI employment.
The Bureau of Labor Statistics projects 33.5% growth in data-scientist employment from 2024 to 2034, 19.7% for computer and information research scientists and 15.8% for software developers. These are occupational projections, not a count of current vacancies or a guarantee that a newly qualified person can quickly enter the market.
Companies also differ greatly in how they define an AI role. One employer's AI engineer may integrate model APIs into a product, while another needs advanced machine learning training, optimised inference, distributed computing or a specialist research background.
The first job-search decision should be which capability you want to sell.
Choose a market that fits your work
San Francisco and Silicon Valley: model companies, startup ecosystems, tooling, AI infrastructure and product engineering. Strong for candidates with practical systems experience and specialised technical portfolios, but competition and living costs matter.
New York: enterprise application buyers, financial services, advertising, media, consulting and regulated industries. Consider applied AI engineering, AI product, data platforms, model risk and governance.
Seattle: cloud infrastructure, distributed systems, enterprise software and technical platform roles. Applicants should know how a model application behaves under real production conditions.
Boston: university research, health technology, robotics and scientific AI. Some research roles require graduate work, publications or strong domain-specific data experience.
Washington, DC and its surroundings: cybersecurity, public sector, federal contracting and AI governance. Certain roles involve citizenship, clearance or contract-specific eligibility requirements.
Austin and other growing markets: software products, enterprise implementation, consulting and startup opportunities. Compare employers individually rather than assuming all AI jobs are concentrated on the coasts.
These city descriptions indicate specialisations, not a claim that one city currently has a particular number of vacancies.
Six role families to distinguish
AI application engineer: integrates models into workflows, measures accuracy, sets up retrieval, monitors quality and handles security or permissions.
ML engineer: builds reliable modelling and data pipelines, trains and deploys systems, and maintains evaluation and monitoring.
Research scientist: develops new methods or models; advanced mathematics, experimentation and published work may be central.
Research engineer: translates research into working implementations and repeatable experiments, often demanding strong software, infrastructure and modelling knowledge.
AI product manager: sets problem definitions, user research, measurable success criteria and product decisions.
AI governance, safety or security specialist: tests system behaviour, risk controls, access, documentation and incident processes. This route can draw on non-ML professional backgrounds.
An internship or transition into AI may require a different portfolio from a senior research position.
Salary evidence without misleading averages
US AI-related compensation varies sharply across role, city, sector, experience and employer type. An advertised base salary is not interchangeable with total compensation including equity and bonus.
Begin with BLS occupational wage data as a broad labour-market reference, then compare current role-specific offers and their location requirements. Equity at a startup has different risk characteristics from a salary at a large corporation. Seniority and domain matter.
Avoid articles claiming one average 'AI salary in America' without naming the occupation, methodology and date. For job applications, calculate how cost of living, taxes and benefits change the real comparison.
A portfolio that demonstrates production readiness
A useful project might be an evidence-backed retrieval system for public documents, a classification service using open data or a model evaluation harness.
Start with a simple baseline: a search index, a rules-based approach or a conventional machine learning model. Then test whether the AI version improves an actual user outcome.
Show error analysis, evaluation data, test coverage, logging, latency, cost per transaction and failure-handling. If you use a hosted model, explain data handling and third-party dependency.
For data-science roles, a reliable experimental design may matter more than a flashy interface. For research engineering, show reproducibility and performance trade-offs. For product roles, show what you chose not to build.
How to approach hiring and networking
Read ten current role descriptions and make a skills matrix. Notice whether the position expects MLOps, cloud, mathematical research, business integration or model evaluation. Tailor evidence to that emphasis rather than claiming every AI buzzword.
A strong outreach message is specific: identify a product or research challenge, share a relevant project and explain one observation about the company's work. Avoid generic messages sent to hundreds of people.
Industry conferences can help with peer learning and finding practitioners, but paid event attendance is not a requirement for employment. Online communities, open-source contributions, professional groups and local meetups can also be effective.
The first month of an AI job-search improvement plan
Week 1: define one main role and a backup option. Select cities or remote eligibility realistically.
Week 2: create or improve one portfolio project with a testable question and clear outcome.
Week 3: obtain feedback from a professional working in the same discipline. Fix reproducibility, clarity and evidence gaps.
Week 4: apply to suitable roles and track which qualifications repeatedly appear in interviews and rejection feedback.
Review the quality of applications, not only the quantity sent.
How Women in AI fits
Our USA AI conference guide and AI Research Scientist Career Guide can help connect professional learning with events and specialist job paths.
The Women in AI Talent Network of 50,000 members welcomes requests from people exploring relevant opportunities. It does not guarantee interviews or placement. Our Summit takes place in London, and for US-based job seekers, potential virtual participation may be more practical than travelling internationally. Free Job Seeker Ticket availability is under planning and not yet confirmed.
Research note: BLS figures above describe particular occupations and forward-looking projections. We do not claim they represent every AI job or predict any individual's hiring outcome.
Research sources
- U.S. Bureau of Labor Statistics, AI, IT and employment projections 2024–2034
- BLS, Data Scientists Occupational Outlook
- BLS, Computer and Information Research Scientists Occupational Outlook
- Interface, Talent In, Talent Out 2026 AI workforce research