What Do You Need to Land an AI/ML Job? Skills by Role

AI/ML covers several roles with different skill needs. See what each requires, which projects to build, and how to start from where you are now.

You open an AI job post and find a huge list of skills: Python, PyTorch, Kubernetes, LLMs, RAG, statistics, cloud, all for one role, and it feels like years of study before you can apply. In reality, AI/ML jobs cover several roles that look very different, and many job posts list what they would like, not what every candidate must have. Once you know which role you are aiming for, the path gets much shorter.

What roles are there in AI/ML?

• Data scientist: frames business questions, analyses data, builds predictive models and communicates results to decision makers.
• Machine learning engineer: puts models into production systems and keeps them fast, stable and scalable.
• AI engineer or LLM application engineer: builds applications on large language models via APIs, such as a chatbot that uses RAG to pull information from company documents.
• MLOps engineer: runs the automated systems that train, test, release and monitor models.
• Roles that use AI as a tool: such as AI product manager or automation specialist, who need to understand AI without building models themselves.

The foundation most technical roles need

Python and SQL

Handle data fluently in Python and pull data from databases with SQL yourself. These two appear in almost every job post in this field.

Statistics and ML principles

Understand train/test splits, overfitting, metrics such as precision and recall, and when a simple model is good enough.

Software engineering skills

Use Git, write code others can read, write basic tests, and understand how APIs are called. These separate people who can work in a notebook from people who can ship to production.

Communication

Explain model results to non-technical people, including being honest about limitations.

Role-specific skills

• Data scientist: experimentation such as A/B tests, feature engineering, storytelling with data.
• ML engineer: frameworks such as PyTorch or scikit-learn, serving models through APIs, Docker and cloud basics.
• AI engineer: prompt design, RAG, function or tool calling, evaluating answer quality and controlling model costs.
• MLOps: CI/CD, monitoring models after release, versioning data and models.

See more on the path in The AI engineer career path in Thailand.

Projects that get you interviews

Employers want to see that you can solve real problems, not just follow tutorials. Good projects:

• Start from a clear problem, such as automatically categorising Thai-language customer complaints.
• Use data other than the popular sample datasets everyone uses.
• Measure results and explain where the model fails.
• Include a README explaining how to run it, and a demo to try if possible.

More ideas in 10 AI portfolio projects employers notice.

You do not need to know everything in the job post, but you must be able to prove you can do what you say you can.

Start from your current field

• Data analyst: building from SQL and analysis toward predictive models and Python is the closest route to data scientist.
• Software developer: you already have engineering foundations. Add ML principles and LLM app building to move toward AI engineer or ML engineer.
• Business professionals: domain knowledge in finance, manufacturing or HR is a strength. Start with AI-assisted work in your current field.

Cautions

• Do not list a skill on your resume just because you watched a tutorial. Interviews in this field tend to go deep.
• Certificates help show you have studied, but they do not replace real projects.
• AI tools change quickly. Invest in durable foundations while keeping up with new tools.
• Be careful with personal or company data in public projects.

How to start

1. Choose one target role from the list above.
2. Collect ten real job posts and count the skills that appear most often.
3. Compare them with what you have and pick the two or three most important gaps.
4. Learn them while building one project that uses those skills.
5. Practise explaining the project and answering technical questions, then start applying.

Start counting skills from real job posts in job search, fill gaps with partner courses, and prepare for interviews with Common AI interview questions.