is the new “hot” role in the technology industry, and many people are eager to get this job.
I see a lot of posts online about how to become an AI engineer in a few months.
Let me be clear: anyone who tells you that you can become an AI engineer in 6 months is selling you a dream.
It actually takes longer, but that doesn’t mean you can’t speed up the process.
For the uninitiated, I’m Egor. I work as a machine learning engineer and am also a career coach for people working in data, AI, and machine learning.
I’ve seen firsthand what works and what is just a waste of time.
Let’s get started!
There is a lot of confusion online, so let’s clarify what exactly an AI engineer is.
The main differences are explained in another article, but in short, an AI engineer is a software engineer who specializes in using and integrating basic GenAI models such as Claude, GPT, and BERT.
They don’t “build” these models from scratch like data scientists or machine learning engineers. Rather, use them to serve a specific purpose.
For example, you can embed a chatbot on your shopping website to help customers find what they’re looking for faster, or add a coding assistant like Cursor to your IDE.
Since AI engineers are professional software engineers, they must understand basic software engineering practices and have deep knowledge of AI systems.
This skill set is rare but in high demand today due to the hype around AI. So, unsurprisingly, salaries for AI engineers are very high, with many companies paying around $200,000 to $300,000, according to levels.fyi.
As you can see, this is a very attractive career with great growth potential. Now, let’s take a look at how we can do this in detail.
Unfortunately, the reality is that it is very difficult to break into AI engineering with zero experience.
This is because the profession requires substantial expertise across data, machine learning, software engineering, and of course AI.
Therefore, you should be a data scientist or software engineer for at least a year before you consider turning into an AI engineer.
Whether you become a data scientist or a software engineer is up to you and your background.
However, I personally recommend starting as a software engineer as it is more closely related to AI engineering roles.
You don’t have to take my word for it either. Greg Brockman, OpenAI CTO, agrees that it’s better to become a software engineer first and then develop your AI/ML knowledge.
Software engineers should strive to learn the tools and technologies needed to become an AI engineer. These include:
- python — The entire AI/ML ecosystem is built on Python, so you should be able to write stable production code in this language.
- SQL —AI revolves around data, and SQL is the language of data.
- Software development tools — You need to know things like gitFor version control, zsh/bashUnderstand the basics and how to create and use API .
- System design technology — The AI system you eventually build will need to scale and will likely be deployed on a cloud platform such as: AWS, azur , GCPusing a tool like dockerand Kubernetes.
resource
timeline
Your schedule will depend on how long it will take you to land a software engineering or data science job.
Realistically, if you have a STEM background, solid knowledge, and put your mind to it, you can land these roles within about six months.
You should then stay in this role for about a year to ensure you have the basics before attempting to move into AI engineering.
There are many guides online on how to get into software engineering. We also have some roadmaps for becoming a data scientist that you might want to check out.
Alongside your full-time job as a software engineer, you need to improve your skills in AI/ML fundamentals to ensure your journey progresses quickly.
You certainly don’t need a PhD-level understanding of mathematics, as you won’t be building these models from scratch, but they will give you the background information to dig deeper into more advanced topics at a later date.
Here’s what you need to know:
- basics of mathematics —A solid overview of statistics, probability, linear algebra, and calculus will help you understand what’s going on under the hood.
- supervised learning —Understand how basic algorithms such as linear regression, decision trees, and support vector machines work.
- unsupervised learning —Understand how basic algorithms such as K-means and K-Nearest Neighbors work.
- Neural network —These are the backbone of LLM, and a better understanding of topics like backpropagation, vanishing gradients, and activation functions will help you debug AI models faster in the future.
- LLM Basics —Although you won’t be building an LLM from scratch, it’s good to have some knowledge of how it works, as you’ll be working with it every day. You will need to learn about areas such as transformers, autoencoders, tokenization, and embedding.
resource
timeline
Your ability to learn the basics will depend on exactly how many hours you study while working as a data scientist/software engineer.
We recommend incorporating these concepts into your daily work as much as possible.
If you study all of this outside of work hours, expect it to take 3-6 months if you apply on your own.
At this point, it’s time to dig deeper into specific concepts and ideas that you use as an AI engineer in the real world.
This field is rapidly evolving and there is something new to learn every month. Here we list the most important and timeless basics.
- AI API— Services like OpenAI API Integrate powerful models without having to build them yourself. This is the fastest way to start building real applications with AI capabilities.
- rapid engineering— Learning how to communicate effectively with AI models is an important skill. Properly crafted prompts can significantly improve model output and are essential for consistent results.
- Search extension generation (lag)— Understand how to connect LLM to external databases, such as: pine coneUse relevant information to improve the accuracy of your AI model’s responses.
- Model Context Protocol (MCP) — A standardized way to connect AI models to external applications such as files, servers, and other apps.
- rung chain—This is the perfect package for working with AI models in Python. We provide all the architecture you need to build and connect your LLM seamlessly.
- Fine adjustment—Understand how to improve the performance of AI models by training them on specific data to improve their response and output for specific use cases.
resource
timeline
Learning these concepts takes slightly less time than learning the basics of AI/ML because there is less material to cover.
Expect it to take about 2-3 months to learn everything to a good standard.
There’s a lot of confusion about what kind of projects you should build to get a job in AI engineering.
Simply put, the best projects are those that are intrinsically motivating for you and that also benefit some kind of end user or client.
The general steps are as follows:
- idea— Brainstorm ideas and topics that are personal to you and problems you want to solve. This should be derived from your own thinking and research. Don’t look it up online or ask people like me for project ideas. Everything I give you will soon become a bad project for you.
- data— Find new and exciting data using public APIs, government websites, web scraping, and more. We want to replicate the messy data we encounter in the real world.
- expand— Must demonstrate the ability to deploy AI systems end-to-end. This includes data storage, data cleaning, model connectivity, and even front-end integration via API or web app. It should be as consistent as possible with the work you do as a full-time AI engineer.
- document— If you don’t tell people about your project, no one will know. Post on LinkedIn and write blog posts to add to your portfolio. Your project should have a clear and well-organized README on GitHub so people can test it themselves. Please share your work as much as possible. This will increase your chances of being noticed by potential employers.
timeline
Creating great projects and building a strong portfolio takes time. Ideally, you should build two top-level projects, which will take about three months in total. This assumes you can spend 1 hour per day building these.
This could be a whole post by itself, but I’ll give you an 80/20 overview of what to do.
resume
To create a great resume, make sure it’s all about AI engineering.
- Maximize your technical skills with tools and technologies relevant to AI Engineering roles.
- Visualize your project clearly using metrics, numbers, and especially financial impact.
- Keep it simple. Neutral colors, single column, easy to read font, and only 1 page.
- Please list any relevant experience as a software engineer or data scientist.
The complete article on how to create a great resume can be found below. Also, check out the ready-made templates you can use.
Make it clear on your LinkedIn profile that you’re pursuing an AI engineering role.
- Please include “AI Engineer” in the heading. Please do not include “aspirations”. For example, who would want to hire an “aspiring” dentist?
- Include keywords throughout the “About Me” and “Experience” sections, but add them organically and don’t create paragraphs.
- Beautify your profile with clear photos and great-looking banners. This makes a bigger difference than you think.
Introductions and networking
Most people think that in order to stand out and get a job, they need to start a bunch of projects and take endless courses.
It’s a complete waste of time.
Referrals are the golden ticket for tech jobs.
Research shows that referrals make up 7% of applications but 40% of all hires. If you are referred, your chances of landing your dream job are almost six times greater.
The leverage is crazy.
The way to get referrals is actually quite simple and all it takes is some confidence on your part.
- Look for companies that are recruiting AI engineers or companies that you would like to work for.
- Browse employees on LinkedIn and find people like you. This could be someone from the same university and background, ideally also an AI engineer.
- Connect and send a DM with what you like about their profile, what they’ve been up to, and anything else personal. Don’t ask for an introduction in your first message.
- Chat with them and ask them questions about their work, projects, and other cool things they’re doing.
- After a few messages, ask for introductions and feedback on your resume.
The process is very simple, but the problem is that people are too scared to do it.
But I’ve never had a bad experience because you always compliment them and lead with open words.
People love to talk about themselves, so all you have to do is appear friendly and show that you’re interested in them.
timeline
Getting a job varies and sometimes depends on luck. However, if you actually get referrals and don’t get distracted from your projects or courses, this should take you 6 months.
So becoming an AI engineer will optimistically take around two years, but you’ll also need to get a job as a software engineer or data scientist first.
It may seem like a long time, but these roles are highly skilled and pay exorbitant salaries. You can’t just take a few courses and expect to get there right away.
If after reading this article you really want to become an AI engineer, that’s great.
However, as I mentioned earlier, you need to become a data scientist first. Luckily, in one of my previous articles, I wrote exactly the steps I would follow if I became a data scientist again.
See you there!
Join our free newsletter. Every week I share tips, insights, and advice from my experience as a data scientist and machine learning engineer. Additionally, as a subscriber, my Free resume template!
dishing the data
Weekly emails to help you land your first job in data science or machine learningnewsletter.egorhowell.com
