My 6 Secret Tips for Getting an ML Job in 2025

Getting a job in the field of Machine Learning in 2025 seems almost impossible, at least if you don't know what you're doing. I myself managed to become an AI researcher in one of the best AI startups in the world. But to get there, I repeated the same mistakes again and again until I myself learned that there are different techniques to get a job in the field of Machine Learning. In this video, I will therefore share my 6 secret tips which, I hope, will show you that there are different ways to get a job in the field of Machine Learning in 2025. Ask yourself the following question: How do you demonstrate your skills? How do you demonstrate that you can bring value to a project? For example, one of my friends once worked on a personal project with the library of transformation of the faces of the cuddly and he remarked that at a certain point, It was unnecessarily or unusually slow and took up a lot of memory. So he dug deeper into the library of Hugging Face Transformers and found something that looked like a bug in that official library itself. He's done a lot of tests to see if he's a bug. But when he saw that his corrections were constantly improving his performance, he had something to share with Hugging Face. He therefore made a request for extraction, made the complete game of round-trip with one of the employees of Hugging Face and finally got his request accepted. Then a little time passed and look at this face who offered him a job as an ML engineer. Now, the corrective he found was certainly not easy to find. But it was a real example of a story of a person I really know and you don't need to jump straight into such difficult projects. There are many small open source projects to which it is easier to contribute. Researchers from universities and even companies have smaller research projects that are open source to which you can contribute. I have personally heard of some stories of this kind, but nobody really talks actively about this approach as being valid for being directly hired or at least increasing your chances of being hired. This presents very few drawbacks. You learn a lot and, as a secondary advantage, you show your work, show your skills, contribute to the community and you will be ideally employed. But now, the question is perhaps, where do you find these interesting projects? There are many interesting projects, but people forget two important facts. They must know how to look for them and they must look for them actively and not hope that they will come to them by magic. I'm not talking about beginner-level projects you do when you start. I am talking about projects that attempt to solve real open problems and finding yourself as a beginner or even an intermediate is a real challenge. When I did my master's, I didn't specifically follow many courses in machine learning, but I worked on many projects and these projects were conceived by people who were before me, doctoral students or post-doctoral students. They knew what the challenges were in machine learning and I could benefit from them. Now, not all of these projects will be big projects, but you really only need one or two that will really increase your chances in your career. So make sure you actively search for projects at the university and do several of them. But what if you want to go further and work on projects that are directly intended for a certain company? Well, you must then expand your search space. Nowadays, there are many interesting open source projects run by researchers from the best universities, startups and even large companies, to which you can contribute. A fairly simple way of finding interesting projects is to search for articles published by these companies. For each article, there is a good chance that it has an open source library that you might want to play with. But you may be wondering if there are other ways? Well, what I also found very useful are the e-mail newsletters and the follow-up of interesting researchers and engineers on LinkedIn. Two newsletters I can recommend are called TLDDR and they have the same track TLDDR and Distinct and a newsletter and a website called crackedengineers.com. TLDDR generally makes interesting technological updates, but also covers a lot of AI and often also shares interesting new AI references. Cracked Engineers, on the other hand, is also a newsletter but a website which may be quite promising in the future. This is a new project by Alexa Gordik that lists interesting startups and their lists to which you can apply. And that's another little bonus tip I have for you. Do not look only at the big tech companies. I work myself in a startup, don't you? I may be a little biased, but startups offer you many opportunities for growth far more than you could get in one of these big tech companies. And in general, you can consider them as easier to access. Large technological companies often require a lot of prior experience or simply to be in a large university. And if you want to jump into AI and start your career, you have to start somewhere. But again, don't consider AI startups as a simple stepping stone to real cool companies. They are literally cool companies themselves. So yeah, look at the AI startups. Anyway, what I also mentioned is that I'm some cool AI researchers and engineers on LinkedIn. And every time they work on a cool project or their company has a cool project that's published or they just like a cool project from another person in another company, you'll suddenly see how many projects exist to which you can potentially contribute. The university researchers in particular are often understaffed and appreciate any form of help. Speaking of cool researchers, how do you really find these people cool? The importance of knowing the right people and how to find them is never really taught. It's always enough to learn to code and to insert difficult random skills here. And these are certainly the minimum requirements at a certain level. But I can't stress enough how knowing the right people has expanded my career options. When I finish most of my university studies, I want to do a stage and then find a cool professor with whom to do my thesis. So I read a lot of articles and looked for people who worked on things similar to what I had worked on and things I want to work on. And then I wrote some cold emails. Yes, yes, I said it. Cold e-mails can also be a valid option. In fact, I sent an e-mail to a Google researcher DeepMind and I was invited to an interview with him. Now, even though I didn't get the internship at the end, I used exactly the same technique or approach to finally get my role as student-researcher with this incredible professor that I had and I was able to then write my thesis. And thanks to this professor, I suddenly had a new contact with a Google DeepMind researcher with whom I was able to join the project. I cannot stress enough the importance of really knowing the people who are in front of you and who may have an existing network from which you can benefit. It is already a good starting point to get to know different researchers or simply people you admire and with whom you can imagine working in the future. Try following different researchers and engineers in AI on LinkedIn or X and they will show you many more interesting people. Even if you don't succeed right away with cold emails, which probably won't be the case, it's still a very good starting point because you have an objective to achieve, but there are many other ways to find ML friends that could be even easier. It's maybe something you could consider networking. And before you skip this game, listen to me. A few years ago, I took part in a networking event organised by Waits & Bias here in Berlin. And I was there, with my introverted self, sitting in the corner of the row just to make sure that only one person could sit next to me, and not two people. I listened to the presenters. In the end, I didn't speak to any of the other participants. But what I did was I talked to the presenters, and all of a sudden, I was a little student who didn't know anyone in the field of ML. I had three relationships with people in real ML companies. I could introduce myself and ask for ways in which I could contribute or how I could eventually get a job at the end. And what is it about hackathons? Hackathons are a formidable and fun event. You learn a lot. You have a fairly easy time getting in touch with people and you can again show your skills. But the most fruitful way, but also the most advanced, is probably to attend conferences on machine learning. For this one, you're not obligated to publish an article, but if you can afford it, you can just go and you'll be exposed to a ton of incredible researchers. The companies will be there and they will have their stand. You can introduce yourself, impose yourself in the lives of others and simply create connections. So if you are confident in yourself and your skills, do not hesitate to ask people what they are working on, if they need help, to present yourself as a valuable asset or a help for their project, and to create links. It would probably be one of the best ways to make friends in person. But if this is not an option for you and you are at university, do not hesitate to use the people who are before you. Maybe some excellent graduates, some good doctoral students, some post-doctoral students or even some professors. I've already shared my experience with a nice professor who had his own network and how I could benefit from it. And by joining a project with one of these doctoral students, I even managed to co-write an article which was published at a high-level conference. Again, I would like to make it clear that nothing guarantees that the first project you participate in or the first person you meet will bring you the same benefits. It's a process and you have to work on several attempts and you really only need one or two attempts to open up a lot of doors. Now, if even that is not accessible to you, or if you think that the people at your university are not as useful as you would have hoped, you are still in a pretty decent situation thanks to the Internet. If you really have no chance or limited chances of finding good friends in your university, in your city, in your country, you simply have to expand to the whole world. Once again, the purpose of this video is to show you many different ways to actively show your skills and to find people who value these skills and you as a person. Being part of a community has many different advantages. You will get to know different projects on which different people work. You can find friends with whom you can work and you can even find people who appreciate you as a person and your skills and even if they cannot hire you directly, they can perhaps give you references. And references are a super powerful tool for getting jobs in a company. But it's really important not to just spam random people on the Internet and ask for references that won't work. And people, even if they tell you they'll give you a reference, won't help you any more or invest the time necessary in the company to get you an interview. So yes, there are many advantages to joining such a community. And if you're looking for specific recommendations, I'll list some Discord channels right here on the screen. You can find regular reading groups where you can learn new things, learn more about new researchers and new engineers and interesting projects, or you can also attend guest conferences of people who come from start-ups, and all of a sudden, you know new start-ups to which you can potentially contribute, start-ups or open source projects. So yeah, there really are no disadvantages to that. So, if you actively seek opportunities and increase your research radius, you'll see how many arise out of nowhere. But what if there was a way to let people find you instead of you finding them? One of my favorite stories is that of a friend of mine who was doing his doctorate while I was a graduate student. He did research in a rather specialized field and every time he learned something new or he did a new project that he wanted to publish, share or talk about or every time he read an interesting article, he wrote a blog post or tweeted something and simply shared his knowledge. From time to time, he received comments and one day, one of these people was a Google DeepMind researcher who loved his work. He was getting along well and this researcher from Google DeepMind simply helped him land a position at Google DeepMind. In short, what he was doing was learning to the public and sharing his knowledge, which ultimately helped him land a job at Google DeepMind. In fact, because of his experience, he also recommended that I continue to make YouTube videos. Now, you don't need to make YouTube videos to get a job or relationships. And it took me about two and a half years to really make a profit. And after two and a half years of creating YouTube videos, I've established contacts with people from Google, Meta, Nvidia, etc. That's really cool, but it takes a lot of time. I mentioned intentionally that it took me two and a half years. Now, again, you're not obligated to make YouTube videos, but think about writing blog posts about the projects you do. Think about tweeting or writing messages on LinkedIn. Simply share your experience, make yourself known and make sure people are interested in you. Every advice in this video is a advice I wish I had. I've already received some advice, but they take time to implement. All these tips that I've shared are things you can try, but you're certainly not obligated to do them all. Find something that works for you. Go step by step, but make deliberate and intelligent steps. I could share many more things here to help you make smarter steps, but if you want to know more, I recommend you watch this video next, where a real ML engineer at Google shares his tips. I'll see you later.