How to Get Your First AI Engineering Job (skills, projects, resumes, and more)

So, you want to break into AI engineering, but every job posting, even an entry- level one, requires years of experience. This puts you in a pickle if

experience. This puts you in a pickle if you're new to the field. You can't get experience without experience. So, what

are you supposed to do? That's what

we're breaking down today. We're going

to talk about what skills you actually need to learn and the best way to learn them, how to make standout portfolio projects, and an alternative to just applying to jobs on LinkedIn and never hearing back. If you're new here, I'm

hearing back. If you're new here, I'm Marina. I work in applied machine

Marina. I work in applied machine learning at Amazon and I'm also a one-on-one career coach for people looking to break into AI and machine learning. I've seen firsthand what works

learning. I've seen firsthand what works and what's just a waste of time. Let's

get started. First, let's clarify what an AI engineer actually does because there's a lot of confusion on this point. Often when someone says AI

point. Often when someone says AI engineering, they're really describing data science or traditional ML engineering, which are actually quite different. I have a whole video breaking

different. I have a whole video breaking down the differences between these roles. But the main thing you need to

roles. But the main thing you need to know is an AI engineer is not someone who trains models from scratch like a data scientist or machine learning engineer would. Instead, AI engineers

engineer would. Instead, AI engineers build applications on top of pre-trained foundation models like GPT5 or Llama.

They focus on model adaptation through prompt engineering, retrieval, augmented generation, fine-tuning, and AI agents.

This is a specialized role that builds on software engineering skills. You're

not creating new AI models from scratch.

You're taking AI capabilities that already exist and turning them into useful, reliable applications that solve real problems. These roles are really in demand these days, often commanding salaries of $200 to $300,000 a year in

the US. And there's a shortage of people

the US. And there's a shortage of people with the relevant skills. Speaking of

skills, what skills do you actually need to get a job in this field? I like to break this down into three tiers.

Foundational skills, core AI engineering skills, and advanced techniques. Before

you can dive into AI specific technologies, you need some basics.

First, strong Python programming is pretty much non-negotiable in this field. You need to be able to write

field. You need to be able to write production level code, not just code in a notebook. You'll also need basic

a notebook. You'll also need basic software development concepts like Git for version control, command line basics, and understanding how to work with APIs. And while machine learning

with APIs. And while machine learning isn't the focus for AI engineers, understanding fundamental machine learning concepts will give you the vocabulary you need to understand more advanced topics. Things like the

advanced topics. Things like the difference between supervised and unsupervised learning, model evaluation metrics, and concepts like overfitting and underfitting. Once you have those

and underfitting. Once you have those basics down, here's where AI engineering really starts. First up is learning to

really starts. First up is learning to use AI APIs. Services like OpenAI's API let you integrate powerful models without needing to build them yourself.

You need to be able to experiment effectively with different pre-trained models so you can pick the right one for your task and measure its performance.

Next, understanding prompt engineering is core to the daily work. Then you'll

want to dive into building rag applications. This involves connecting

applications. This involves connecting AI models to your own data sources using vector databases and embedding techniques. This allows them to be able

techniques. This allows them to be able to answer queries based on your specific information. Today, many teams are

information. Today, many teams are building a lot with AI agents as well.

So, understanding how these systems work will be really important. And finally,

deployment and infrastructure. You'll

need to learn containerization with Docker and cloud deployment on platforms like AWS, GCP, or Azure. You'll need to understand system architecture and things like monitoring and logging.

There are also some advanced techniques you may not need to get your first job, but learning these will make you stand out and help you grow in your career.

Some things to think about could be advanced rag techniques like implementing more sophisticated chunking strategies, optimizing embedding techniques, and understanding the trade-offs between different retrieval methods, mastering fine-tuning

techniques with Laura, and making intelligent model selection decisions based on trade-offs between things like cost, performance, and licensing, and security, privacy, and ethics to implement safeguards against attacks like prompt injection, ensure privacy,

compliance, and consider the ethical implications of the AI systems you're building. So, that's a lot of skills.

building. So, that's a lot of skills.

The next logical question is how should you actually learn all of this? [music]

When it comes to breaking into tech, there are typically four main paths people take, and the right one really depends on your situation. Let's start

with self-study. This costs basically nothing to maybe $1,000 for courses and books. The timeline is super flexible

books. The timeline is super flexible cuz you can control how fast you learn, and the opportunity cost is minimal, especially if you're learning while you're maintaining your current job. But

the challenge is that it requires significant self-discipline and motivation since you'll need to structure your own learning journey. and

you'll need to build your network independently through meetups, conferences, open source projects, online communities, and things like direct LinkedIn outreach. That takes a lot of effort. If you want more structure mentorship and accountability, boot camps and

certificate programs are another option.

Boot camps cost between $5 to $20,000 and take 3 to 12 months. Many offer

part-time options let you keep your day job. The best boot camps have strong

job. The best boot camps have strong partnerships with companies who specifically look to them for talent.

You get industry guest speakers or instructors, which can ensure you're actually learning what we really use on the job. and you'll of course make

the job. and you'll of course make connections with your cohort. Speaking

of programs that stay current with industry needs, today's video is sponsored by SimplyLearn and I'm excited to tell you about their applied generative AI specialization program which is delivered by SimplyLearn in partnership with Purdue University

online. What caught my attention here is

online. What caught my attention here is how practical this program is. Rather

than just focus on theory, you'll use the actual tools we use on the job like Lang Chain, OpenAI, and Hugging Face.

During the program, you build seven industry projects including rag applications and agentic AI systems. The program is specifically designed for working professionals and runs 16 weeks with live online classes, so you're

learning from instructors in real time, not just watching pre-recorded videos.

You get master classes from Purdue faculty, plus career services like resume help and mock interviews. Whether

you're a software developer, data analyst, or even transitioning from a completely different field, the curriculum scales from Python basics all the way through advanced topics like fine-tuning and agent systems. You can check out the full curriculum and

student reviews at the link in the description. Simply Learned has trained

description. Simply Learned has trained thousands of learners who are now working in AI roles. A more involved option is to get a master's degree. M's

programs range from $10,000 at the low end for Georgia Tech's MSCS program all the way to over $100,000 for private elite universities. You're looking at 1

elite universities. You're looking at 1 to two years full-time or 2 to three years part-time, which means the opportunity cost can be substantial, especially if you're studying full-time and giving up employment. My main

criticism of M's programs, though, is that the curriculum is usually really outdated and overly focused on theory.

So you spend a ton of time and money just to be learning outdated approaches.

But the benefit is that the credential means you won't automatically be screened out by automated resume screening tools which is a very real thing. Occasionally people also ask

thing. Occasionally people also ask about the relevance of PhDs. While these

are generally funded so you make a little salary. The timeline is 4 to 6

little salary. The timeline is 4 to 6 years. So the opportunity cost is huge.

years. So the opportunity cost is huge.

For AI engineering where you're not actually working on the models themselves, this is not a good use of your time and money. So given these four options, which makes the most sense for you? For AI engineers specifically, a

you? For AI engineers specifically, a technical bachelor's plus self-study is sufficient if you're targeting startups or mid-size companies and can demonstrate solid skills with challenging projects. Choose a boot camp

challenging projects. Choose a boot camp if you already have a technical background and you want to quickly pivot into AI engineering or if you need structure and accountability. Look for

programs with strong job placement rates and industry partnerships. Get a masters if you're aiming for fangle level companies and you're willing to do it part-time while you work. Focus on

programs with strong practical components and be ready to do extra self-study. No matter what path you

self-study. No matter what path you choose, you'll almost definitely need to do additional study on your own and build portfolio projects. So, let's talk about what makes a good project next.

There are lots of different ways to approach building your portfolio, but unfortunately, many people waste a lot of time here building things that won't actually help them with their career.

Let's talk about the common approaches from least to most effective. Starting

at least effective, we have followalong tutorials, which are great for learning the basics, but they won't actually land you a job. Nothing you build in a Corsera or Udemy course counts at the professional level, unfortunately.

Certificate capstones and Kaggle competitions are better, but still not very unique and often have too much handholding. Self-motivated projects

handholding. Self-motivated projects where you come up with your own problem and source your own data are starting to get useful. You'll want self-motivated

get useful. You'll want self-motivated projects that solve real problems using AI engineering tools and applications like Rag, chat bots, or agent systems. Even better is creating real projects for real clients. even if you're working

for free or volunteering. Working with

real data under real constraints shows you understand industry expectations and gives you the opportunity to legitimately put AI engineer on your resume. So, if you're building projects

resume. So, if you're building projects for yourself or for someone else, here's a framework to build something that actually stands out. Choose a topic that genuinely excites you so you have some domain knowledge and can showcase your

personality. Avoid pre-cleaned common

personality. Avoid pre-cleaned common data sets. Use public APIs, web scrape,

data sets. Use public APIs, web scrape, find unusual government data sources or niche industry surveys. generate your

own data through experiments or combine multiple data sets in novel ways. For AI

engineers especially, you need to show you can build complete solutions, not just call an API one time. This means

creating end toend pipelines that include data collection and storage if that's relevant for your project, possibly data cleaning and pre-processing, model integration, whether that's APIs, fine-tuning or rag,

application deployment, and some kind of user interface or API with relevant monitoring and logging. You're trying to replicate how we actually work in industry as much as you can, but technical skills alone won't get you

hired. Communication is one of the most

hired. Communication is one of the most important skills for AI engineers. Make

your project accessible and impactful by creating a really nice GitHub repo with modular, welldocumented code. Write a

compelling readme that explains why your project does what it does and why that matters. Include clear instructions for

matters. Include clear instructions for setting up and running your project. And

ideally, make some kind of interactive UI. Then share your work as wide as you

UI. Then share your work as wide as you can. Write a blog post and share on

can. Write a blog post and share on LinkedIn Twitter Discord Reddit wherever. You could also create YouTube

wherever. You could also create YouTube videos demonstrating your project, or even present at local meetups or conferences. The more you share your

conferences. The more you share your work, the more likely it is to be seen by potential employers or collaborators.

All right, so now you have the skills and have demonstrated them through real end-to-end projects. It's time to start

end-to-end projects. It's time to start applying to jobs. When it comes to your resume and LinkedIn, structure your materials to lead with relevant content.

Put your skill section first with all your AI engineering technologies highlighted right up top. Make your

project section prominent and put your portfolio website link somewhere where it's clearly visible. For your LinkedIn headline, instead of saying something like student or aspiring AI engineer, use something like AI engineer building

production LLM applications, rag, and finetuning. Write a summary with

finetuning. Write a summary with relevant keywords that positions you as already an AI engineer. All of this can help set you up for success in the traditional job application process. But

unfortunately, if you lack industry experience, you may still get filtered out by automated resume screening systems. So, we need to get creative. At

this point, I always recommend networking, but probably not how you think. It's not just about going to

think. It's not just about going to local meetups and hoping to meet someone, although that's great, too.

We're actually going to proactively build up our network through non-pushy, cold outreach. And before you poo poo

cold outreach. And before you poo poo this, cold emails are how I got my first data science job with no experience. So,

I know it can work. I suggest reaching out to people from two kinds of companies. First, companies that you see

companies. First, companies that you see are hiring AI engineers, of course. And

second, companies that you'd love to work for, whether or not they have open positions right now. If a company is actively hiring, you can reach out to the recruiter or the hiring manager. If

they aren't, anyone on the team doing AI engineering is a good choice. Now,

here's the thing. We are not going to just send a message on LinkedIn asking for a referral or saying that you'd love the job. We're actually not going to

the job. We're actually not going to mention hiring at all. Instead, find

something genuinely interesting that the company is doing. Read their technical blog, watch their YouTube, read their papers on archive. Then send them a message praising their work and asking some kind of detailed, informed question. Why did you take this approach

question. Why did you take this approach versus another approach? What motivated

this decision? Things like that. You're

just trying to compliment them. Ask a

smart question and open a dialogue. If

they respond and you have some friendly, well-informed exchanges, then you can ask for feedback on your resume or a project or ask what skills they're looking for on their team or maybe even directly ask for a referral. I know this

sounds scary, but I can say personally I'm never offended by a well-thoughtout personalized message. Though I do tend

personalized message. Though I do tend to ignore the pretty generic [music] ones. And honestly, the worst outcome is

ones. And honestly, the worst outcome is that nothing will happen. They'll just

ignore you. But the potential upside is enormous. So, it's worth it. All right.

enormous. So, it's worth it. All right.

So, as you can see, there's a lot involved with getting that first job.

Many people ask me what's realistic in terms of a timeline. So, let's talk about that next. As you can see from this list of really technical concepts to learn and projects to build, you're not going to just go from an absolute beginner to a professional AI engineer

in 3 months. Assuming you're working on this part-time and self-studying, here are some rough guidelines for how long each phase will take. Getting the basics down and building your first AI apps will probably take around 6 months. This

assumes you're studying consistently and already have some programming background. If you're starting from

background. If you're starting from absolute zero, add another 6 months.

Becoming comfortable with more advanced concepts will probably be 6 to 12 more months. During this time, you should be

months. During this time, you should be building increasingly complex projects and deepening your understanding of the underlying technologies. Reaching

underlying technologies. Reaching professional level competence could be another 1 to two years. This is where you develop those specialized skills needed for roles at larger companies or more advanced AI applications. Becoming

a senior or lead AI engineer at a top company where you'd really be earning $300,000 or more would be 3 to five additional years. Mastering the full

additional years. Mastering the full spectrum of AI engineering skills and gaining the experience to lead complex projects takes significant time in the field. In total, expect a 3 to 5 year

field. In total, expect a 3 to 5 year journey if you're starting from scratch and working part-time on your learning.

Is that a long time? Yes, but it's also realistic. And remember, you can start

realistic. And remember, you can start building and potentially even working in the field much earlier than that. You

don't have to wait until you've mastered everything to get started. I just say this because those who expect to become AI engineers in just a few months inevitably get discouraged when they discover the field's true complexity.

But those who approach it with realistic expectations and focus on incremental progress often succeed and find themselves in high demand positions that didn't even exist a few years ago. If

you want to dive deeper into any of these topics, I have detailed videos on building standout AI engineering projects, a complete AI engineering roadmap with a comprehensive skills checklist, and interview prep for AI and ML roles. Check those out. The links are

ML roles. Check those out. The links are in the description. And if you want personalized guidance, I offer one-on-one mentoring sessions. That link

is in the description, too. Thank you so much for watching, and I'll see you next time.