All articlesEntrepreneurship

Skills Required for AI Jobs: The Two Lists You Need

EX EPIC Academy·2026-09-28
Skills Required for AI Jobs: The Two Lists You Need

Skills required for artificial intelligence jobs split in two: the stack for building AI and the skills for jobs that use it. Find out which list is yours.

Search for the skills required for artificial intelligence jobs and you get the same answer on every page: Python, statistics, linear algebra, machine learning, deep learning. Read that list and one of two things happens. Either you nod, because you are an engineer, or you quietly decide AI jobs are not for you. Both reactions come from the same mistake. There is not one list. There are two, and most people searching this phrase need the one nobody publishes. ## The short answer: it depends which list you are on List A is the set of skills required to build AI: training, fine-tuning, integrating and running models. List B is the set of skills required for jobs that use AI: marketing, operations, consulting, research, sales and product roles where AI does part of the work and you are accountable for the result. Both lists are growing fast. Unique job postings asking for generative AI skills went from 55 in January 2021 to nearly 10,000 by May 2025, an increase of more than 18,000%. And it is not only senior roles. A survey of employers by the National Association of Colleges and Employers found more than one-third of entry-level jobs now require AI skills, nearly triple the share six months earlier. The question is not whether you need AI skills. It is which ones. ## Why every skills list you have read feels wrong The pages that rank for this search are mostly written by universities and course platforms selling technical programs, so they describe the technical job. The San Diego list is openly aimed at people "looking specifically for tech roles." That is honest, and it is also a small slice of the market. Lightcast's analysis of more than 1.3 billion job postings, reported by Dataquest, found 51% of postings demanding AI skills sit outside IT and computer science, and generative AI roles in non-tech industries grew 800% since 2022. Half the demand is on List B, and almost all the advice is written for List A. Even Salesforce, which sells AI software to engineers, splits AI skills into two categories: technical skills to build and maintain AI systems, and non-technical AI literacy to work effectively with them. It also reports that organisations name finding people with the skills to implement AI as their third-biggest barrier. The gap is not only a shortage of engineers. If you want the underlying definition first, we covered what AI skills actually are separately. Here we go straight to the lists. ## List A: skills required to build AI This is the list the rest of the internet gives you, and it is correct for its audience. The Johns Hopkins engineering school's version is a fair summary of the core. ### Programming: Python first Python is the entry ticket. Johns Hopkins names Python, R and Java as the base languages, Python for its ecosystem, R for statistics and visualisation, Java for large systems that need to scale. If you only learn one, learn Python. ### Math: statistics, probability, linear algebra You need enough statistics to know whether a model is actually good, enough probability to reason about uncertainty, and enough linear algebra to understand what a neural network is doing. You do not need a math degree for most roles here. You do need to stop being afraid of the notation. ### Machine learning and deep learning Supervised, unsupervised and reinforcement learning, then neural networks. The San Diego list groups these with natural language processing and computer vision as "core modeling" skills, and maps them to titles like Machine Learning Engineer, Data Scientist and Computer Vision Engineer. ### Data wrangling Models are only as good as the data they eat. Cleaning, organising and preparing datasets is a large share of real AI work, and it is the part beginners skip because it is not glamorous. ### Production: MLOps, cloud and evaluation The second San Diego bucket is production and infrastructure: data engineering, MLOps and cloud platforms. This is where a notebook becomes a system someone else can use. One split inside List A matters more than any tool. Dataquest separates builders, who train and evaluate models, from integrators, who build products on top of models that already exist. Integrators need the code but not the research math, and Dataquest calls this the most realistic entry point in 2026. Its honest timeline for the full technical route is roughly 8 to 9 months at 5 hours a week. Graduate degrees mostly matter only for frontier research, which is a narrow slice of AI work. ## List B: skills required for jobs that use AI This is the list for the other half of the market. Dataquest calls it the AI-powered professional lane, with titles like AI Product Manager, AI Implementation Specialist, AI Operations and AI Evaluation Specialist. Here are the skills, written as things you can test yourself on. ### Task decomposition Can you take a messy request, say "research the market for this product", and break it into steps an AI system can do well and steps that need a human? This is the skill that separates people who get value from AI from people who get plausible noise. ### Prompting and context Not magic phrases. Giving the model the right input: the brief, the audience, the constraints, an example of good output. Most bad AI output is a bad brief. ### Output judgment The most underrated skill on either list. Can you spot a wrong number, an invented source, a confident paragraph that says nothing? If your job uses AI, your name is on what it produces. Checking is the job. ### Workflow and automation Connecting tools so a task runs repeatedly without you: a form that feeds a sheet that triggers a summary. No code or light code. If you want to move further toward List A from here, see how to go from AI user to AI builder. ### Domain knowledge The part people discount. AI is useful when someone knows which problems are worth solving and whether the output fits the real world. Your field is not a gap to close. It is your advantage. Notice what employers themselves say about this list. When Strada surveyed executives about entry-level hiring, they ranked AI literacy as the least important skill, and critical thinking and communication as the most. That is not a contradiction of the NACE numbers. It means employers assume the tools are learnable and hire for the judgment around them. ## The skill both lists share Put the two lists side by side and one thing is missing from both, because it is not a topic you can study. It is shipping finished work that someone else relies on. The Strada employers ranked real work experience as the strongest hiring signal. A certificate shows you covered a syllabus. A deployed model, a live automation, a report a client actually used, shows you can take a task from start to done when it is messy and nobody is grading. This is the entire design of EX EPIC Academy in Canggu. Founder Julien Uhlig describes it bluntly: participants from 26 nations working on 15 live projects, with no grades and no exams. Participants are placed in operational roles at real ventures including Zero X, Gemino AI and LIV Urban Sanctuary. People can arrive with List A or List B knowledge. The point is to leave with the shared skill, because a live venture does not accept a half-finished draft. ## Which list is yours: a quick self-check Answer three questions honestly. 1. Do you want to write code most of your working day? If yes, you are on List A. Start with Python and one small model you evaluate properly. 2. Is there a field you already know or love? Marketing, finance, healthcare, sustainability, operations. If yes, List B is faster and that field is your edge. 3. Do you want to build products on top of AI without doing research? Then you are an integrator: List B judgment plus the coding half of List A. Whichever you pick, the next step is the same: one finished project, not ten courses. Our guide on how to learn AI skills for jobs walks through the order, and these AI project ideas for beginners give you something to ship this month. If you want to work from anywhere, the AI skills for remote jobs lean even harder on List B. ## How employers check the skills Knowing the list is half the problem. Proving it is the other half. - List A is usually tested directly: coding rounds, take-home model tasks, and increasingly interviews where AI assistants are allowed and you are judged on how you use them.

  • List B is checked through your work. What did you produce with AI, who used it, what changed? A line saying "proficient in ChatGPT" proves nothing. A link to a finished deliverable proves a lot. We covered how to put AI skills on your resume in a way that survives a follow-up question. If you want to build that proof inside real companies rather than on side projects, EX EPIC Academy runs 4 to 6 month on-site internships in Bali across AI and automation, engineering, consulting, content and research tracks. You will not be graded. You will ship. ## FAQ ### Do I need a degree to get an AI job? Not for most AI jobs. A graduate degree matters mainly for frontier research roles. Integrator and AI-using roles hire on evidence of finished work, and many List B roles value your existing field more than a new qualification. ### Is math required for artificial intelligence jobs? For building AI, yes: statistics, probability and linear algebra, to a working level rather than a research level. For jobs that use AI, no. What you need instead is the numeracy to notice when an AI output is wrong. ### Which programming language is best for AI jobs? Python. It has the largest AI library ecosystem and is the default for model work. R helps for statistical analysis and Java appears in large production systems, but neither is where a beginner should start. ### What AI skills do freshers need for entry-level jobs? Freshers need one practical AI workflow they can explain end to end, the judgment to check its output, and at least one finished project to show. Employers treat tool knowledge as learnable and hire for critical thinking, communication and real experience.

Want to build this way instead of reading about it? Email academy@exventure.co.

EX-AI Summit 2026 · 18–20 Nov

Online · Las Palmas · Bali

Three days with the operators building what replaces the old map: academy alumni, portfolio founders and the people who fund them, in one room.

EX Venture Group and its affiliated service companies are incorporated and operate in the United States and Europe only. They do not operate in Indonesia and do not provide services into or from Indonesia. All Bali-based operations, including EX Epic Academy, are solely managed and operated by EPIC Solutions PT.

© 2026 EX Epic Academy · Operated by Epic Solutions Global, Bali, Indonesia

▸ UU PDP No. 27/2022 · UU ITE · Cookie consent required

▸ Cookies & Privacy

We use essential cookies to operate this site and analytics cookies to improve it. In line with Indonesia’s UU PDP (Personal Data Protection Law, UU No. 27 of 2022), your consent is required for non-essential cookies. See our Privacy Policy.