How To Land A Machine Learning Job If You Don’t Have a Degree
We have a crisis on our hands. Just about every company in the world is trying to figure out how to leverage artificial intelligence (AI) to improve their bottom line, but to implement AI in their business operations they require the expertise of an AI professional.
The problem is there are not enough AI professionals around to fill all the available AI job openings.
According to new research from software provider SnapLogic, 93% of US and UK organizations consider AI to be a business priority, but 51% say they don’t have the necessary AI expertise on board to reach their business goals. In fact, the AI skills gap is the main reason for the lag in AI adoption.
One of the reasons for the shortage in AI skills is that, in practice, AI is developing so fast that universities and other training institutions can’t keep up with developing up-to-date curricula. Academic institutions are not delivering enough graduates that are ready to take on AI jobs.
Machine learning engineer – sought-after but too few qualify
Of all the AI jobs out there, machine learning engineer is the most sought after. According to a report from job site Indeed, machine learning engineer is the top job of 2019 requiring AI skills. ML engineers can command an average salary of $146,085. According to Indeed this career showed a growth rate of 344 % in 2018.
So with this growing demand for ML engineers, coupled with a shortage of qualified applicants, many people are asking if it’s possible to enter the field without a formal academic qualification.
Is it possible to become a machine learning engineer without a computer science degree?
If you judge by the success of many startup founders and developers who don’t have computer science majors, it certainly seems possible.
The academic approach to computer science has often come under severe criticism from startup CEOs and observers. Here is an incomplete list of some of the reasons to consider skipping a degree in computer science.
- Computer science departments are years behind in the AI field, which is changing constantly.
- Computer science departments prepare their students for academic or research careers, not for the realities of the job market.
- Students are taught how to design an operating system, but not how to function effectively in a development team.
- For the most part, academic training doesn’t produce top developers – top developers are usually self-taught.
- A degree is expensive. A degree program can set you back tens of thousands of dollars and while you’re studying, you’re not making any money.
- A degree takes many years out of your life, especially if you go for a higher degree. And during this time study debt is piling up.
How to get started in machine learning without a degree
If you have been coding since childhood, chances are that you have what it takes to become a ML engineer.
Joel Beasley, author and host of Modern CTO and one of Entrepreneur’s Top entrepreneurs under 40, is adamant that it’s possible to get a job as an ML engineer without a degree, and he is not alone in his opinion.
In an interview with Engineered Truth, Beasley said what you need is not a CS degree, but the experience and capabilities to actually execute the job.
He suggests if you decide not to get a CS degree, to find something that interests you and see how machine learning can apply to it. He gives music as an example. Say you want to know what note your favorite artist sings most often – figure out how to make that into a ML project.
A personal project like this is one way to gain the experience you need to apply for a job as an ML engineer.
Other ways that you can gain the necessary experience is by participating in hackathons, coding challenges and open source projects. You could even consider doing your own ML research project.
Writing for Towards Data Science, Donal Byrne says your project doesn’t have to be impressive; it must just illustrate your skills and understanding of the topic. Your aim is to put your project on GitHub where it can be seen, evaluated and commented on by anyone.
Attend a hackathon
Hackathon projects can offer a great addition to your portfolio if you don’t have much experience. Hackathons provide a network of like-minded peers who you can learn from and share with. Try to team up with skilled developers so you can learn from them. There are hackathons specifically aimed at machine learning, but if you feel you lack the necessary knowledge, consider events that focus on AI or software development.
Participate in coding challenges
You don’t have to attend a formal coding challenge to benefit from this challenge and improve your coding skills.
You can sign up with HackerRank and participate in one of their regular challenges. Programmers from all over the world come together on the platform to solve problems in a wide range of computer science domains, including algorithms, machine learning, or artificial intelligence, etc. You can even get hired by solving challenges.
Participate in open source projects
Open source projects are very useful for programmers. You gain invaluable experience working on an open source project. You don’t only develop your technical skills; you also learn how to work as part of a team, an ability that is highly valued in the workplace.
Here is a great resource for machine learning open source projects that includes units for beginners.
Tackle a machine learning research project
Machine Learning Mastery suggests that if you are obsessed with a particular concept or machine learning method, you can design your own research program. As the writer notes, anyone can read and internalize research papers, write down their own ideas and design and execute their own experiments. It will not be easy to do outside an institution, but the writer suggests three ways you can practice machine learning research outside of an institution.
You’ve gained experience in machine learning, now what?
When you apply for a job as a machine learning engineer, you will still be faced with the reality that most job ads will stipulate that you need a degree.
Don’t be put off by that, says Beasley. The person who wrote the ad, wrote it like that because he believes that a person with a degree can solve his problem. But you must look beyond that. Ask yourself: why did he write this ad? What is the problem that he is faced with? Consider if you can solve his problem. If you know you have the skills set to make that problem disappear, contact that person and convince him of your skills.
How do you prove your machine learning skills?
This is where Beasley’s other valuable advice comes in.
He suggests sharing your knowledge and pushing that knowledge out on GitHub where you will get eyeballs on your work. On GitHub people can see your code and they can see if your code is solving their problem. When you apply for a job, you can link to GitHub in your application.
Also, get yourself physically near a person that has the skills that you want. So before you apply for a job as a machine learning engineer, apply for a job as a coffee runner for an ML engineer. That way you will be near the experts and hear conversations focused on machine learning. In the process, you learn many things you need to know and no unnecessary info that’s so often part of online classes.
But that’s not all.
The important thing here is relationships. Everything in life happens because of relationships.
If you want to work as an ML engineer for Facebook, find out who the top ML engineers are at Facebook, says Beasley. Then find their blogs, their Twitter accounts and figure out how you can interact with them.
Even better, target a company that the whole world is not trying to work at. Contact the incumbent ML engineers and build a relationship with them. They might be more receptive to you contacting them and at a later stage recommend you for a job. Says Beasley: “Go out there and put yourself in a position where you can build relationships while learning.”
If you can build a good relationship with someone who is brilliant in the field, people will notice your association with that person.
So, aim for a person, not a company. Relationships with people last, while you can always get fired from a company and that relationship is over. The more people you know and the more proactive you are, the more job security you have, says Beasley.
Admittedly, this route to a position as an ML engineer is not for everyone. It will most probably only bear results for someone who is proactive, a self-starter and proficient in coding. And as mentioned before, the best coders are self-taught.
If you are looking for a job in AI, whether in machine learning or another AI field, check out AIJOBS. It’s the best resource online for jobs in AI. Good luck!