New work in the AI era · 8 min read
Careers in artificial intelligence: the full map, technical and not
AI careers range from research scientist to AI trainer, product manager and policy adviser. A map of the main roles, what each needs, and where to start.
The short answer
Careers in artificial intelligence fall into four broad groups. Building AI: research scientists, machine learning engineers, AI engineers and data engineers, who need strong maths or software skills. Running AI: MLOps engineers, AI operations specialists and evaluation and red teaming roles, which keep systems reliable and safe. Applying AI: product managers, solutions and forward deployed engineers, implementation and automation specialists, and conversation designers, who turn models into products and working processes. Governing and explaining AI: governance, ethics, policy, trust and safety, and AI training and education roles. Only the first group strictly needs a computer science or maths background. Most people enter AI by combining AI skills with expertise they already have, rather than starting over.
Key takeaways
- AI careers split into building, running, applying and governing AI.
- Only building roles usually need deep maths or software skills.
- Applying and governing roles reward domain expertise.
- Most people enter by adding AI skills to existing expertise.
On this pageThe four groups of AI careers
The four groups of AI careers
| Group | Example roles | Typical background |
|---|---|---|
| Building AI | Research scientist, ML engineer, AI engineer, data engineer | CS, maths, software |
| Running AI | MLOps, AI operations, evaluation, red teaming | Software, IT, security, QA |
| Applying AI | AI product manager, solutions engineer, implementation, automation | Product, consulting, operations, domain |
| Governing and explaining AI | Governance, ethics, policy, trust and safety, AI training | Law, policy, compliance, teaching, writing |
Building AI
Research scientists invent new methods and usually have PhDs. Machine learning engineers train and optimise models. AI engineers build products on top of existing models, using APIs, retrieval and agents, and have become one of the most hired roles. Data engineers build the pipelines that feed all of it. These roles pay the most and are the most competitive.
Running AI
Once a model is in use, someone has to deploy it, monitor it, test it and fix it. MLOps engineers handle infrastructure; AI operations specialists watch how systems behave; evaluation and red teaming specialists test quality and safety. These roles suit people from software, IT, QA and security.
Applying and governing AI
Most organisations don't build models; they use them. They need product managers who know what to build, solutions and forward deployed engineers who make it work for customers, implementation and automation specialists who change workflows, and governance and policy people who keep it legal and safe. Domain expertise is a big advantage in all of these.
How to pick your group
- If you love maths and code for its own sake: building.
- If you like making systems reliable: running.
- If you like people and problems: applying.
- If you like rules, risk and explanation: governing.
Where to start
Pick one role in one group and do a small piece of its work this week, using free tools. That tells you more than a course. OffLadder's directions include several AI-era roles, each with a 15-minute experiment, and Laddie, OffLadder's coach, can suggest which fits the experience you already have. When you find a post, OffLadder CV shows which must-haves your CV doesn't show yet.
A note on hype
AI careers attract a lot of hype and many courses promising quick routes into six-figure jobs. The real market rewards evidence: things you've built, tested, deployed or governed. Treat any promise of easy entry with caution, and judge courses by what you'll be able to show at the end.
Illustrative example
Aiyana, a hospital pharmacist, wants to work in AI. Instead of retraining as an engineer, she joins her trust's project testing an AI prescribing assistant. She builds evidence in evaluation and governance and moves into a clinical AI safety role.
Test a future job before you train for it
A forecast can tell you which jobs are growing. It can't tell you whether you'd enjoy the work. Laddie, our AI coach, starts from the experience you already have and works through four steps with you:
- Explore: directions your experience carries into, including work that's only now taking shape as AI changes jobs.
- Try: an experiment small enough to finish this week, a small piece of the work itself.
- Learn: note which part you liked and which part you avoided, and keep it as a record of what you tried.
- Apply: when a direction turns into a job post, OffLadder CV shapes your CV for it. What you can't show yet goes back to Laddie as a short lesson, a 15-minute task and a mock interview.
Questions people ask
- What is the best career in AI?
- The one that fits your strengths. Engineers may suit ML engineering; communicators may suit product or solutions roles; people with legal or policy backgrounds may suit governance.
- Can I work in AI without a degree?
- Yes, in many applying and operating roles, such as AI operations, automation, implementation and AI training work. Research and some engineering roles usually expect degrees.
- Which AI jobs are growing fastest?
- AI engineering, roles that deploy AI with customers, AI governance and roles that build and run AI agents have all grown quickly since 2023.
- How long does it take to get into AI?
- From a related technical role, months. From a non-technical role into an applying or governing role, often six months to a year of building evidence. Into research, years of study.
- Do I need to learn Python?
- For building and running roles, yes. For applying roles, it helps. For governing roles, it's optional but useful.
- Are AI jobs only at tech companies?
- No. Banks, hospitals, retailers, governments, manufacturers and law firms all hire for AI roles, often with less competition than AI labs.
- Is it too late to get into AI?
- No. Adoption is still early in most organisations, and new roles keep appearing. What has changed is that employers want evidence of practical work, not just interest.
Exercises are practice, not evidence of career fit or a qualification. How to read our guides and report a correction.
Sources
- Computer and information research scientists: Occupational Outlook Handbook · US Bureau of Labor Statistics
- Future of Jobs Report 2025 · World Economic Forum
How this page is written
Sources support the factual background, not a guarantee that an exercise predicts career fit. Fictional practice material is labelled. Dates record content updates; no independent expert review is implied. More about OffLadder.