This essay is based on a conversation with Girik Malik, an applied scientist at Amazon Web Services. The following has been edited for length and clarity. Business Insider confirmed his employment and academic history.
At Northeastern University AI in 2023 when he received his PhD, the AI job market was now a world away from the massive AI talent war.
After an internship in the summer of 2022, I secured my job as an applied scientist at Amazon. This was before the company carried out employment freezes for the remainder of the year.
Today we were able to see them earning $1 million paydays from companies like Top Flight AI Researcher Meta, who earned their PhD. However, the road to earning a PhD is not easy.
It requires a lot of discipline and motivation. In fact, during one of my semesters, I was caught up in coursework and research so I had little time to eat. I ended up making a simple stew and ate it every week throughout the school term. It was that difficult.
It takes time. PhD candidates can spend about five years trying to complete it. At that time, they could have spent gaining practical experience in the labor world.
Getting a PhD is challenging, but rewarding
There is still a lot of value to gain from completing your PhD. With AI.
Of course, you can try to self-learn about the field through practical training. You can even build some products yourself. However, if you want to shape the future of AI, you will need to earn a PhD.
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
You will go through the PhD process to make your learning process more structured and accurate. Given how competitive AI research is, it is now extremely important.
Just 10 to 15 years ago, small changes or tweaks to the model will give you a variety of results that could potentially land on an honorable diary or invitation to a high-level meeting. That's not the case anymore.
The bar is much higher. More experiments need to be performed and results that are well generalized across a variety of issues must be presented.
It's never too late to get your PhD. With AI
Some people think that AI is progressing very rapidly, so it may be too late to enter the field through a PhD. root. Defenders believe that all AI innovations and breakthroughs could be discovered by the time they graduate.
I disagree with that view. With PhD AI, it will be the same as it will be related in 50 years. This is because AI is ubiquitous in our daily lives. They can be built into almost every code we write and the systems around us. People who can understand AI and work will need more, not fewer.
AI systems are not perfect. Their algorithm may break in the future. When that happens, you will need someone to understand AI at a basic level and understand AI to fix it. It will likely be someone with a PhD. Among them.
It's like owning a car. It can be a totally great car, but you still need to train your mechanics to fix it when it breaks down. I didn't stop teaching mechanical and automotive engineering just because I could make a better car. The same can be said about AI.
Industry internships are important
I highly recommend that you do an internship during your summer vacation during your PhD. the study. This is important even if you intend to stay in academia.
In my case, I did three internships at the Artificial Intelligence Center in Bosch, Microsoft and Amazon. I was able to get a return offer from Microsoft and Amazon.
The biggest part of working in the industry is that you can easily access data so you can reach and calculate it.
In a university setting, you often compete with 5-6 doctors. Students use computing clusters. And even if the algorithm can be executed, it will ultimately be based on a limited dataset running on limited hardware. The algorithm may not work even after scaling.
So make the most of these internships. The internship I did was forced to solve complex problems beyond textbooks.
One of the challenges I faced during my first internship at Bosch was training my neural network on a large dataset. For example, you could use a large dataset with over 2 million videos. It was much larger than I was used to in college.
Addressing that issue gave me the skills to solve other infrastructure-related issues I encountered later in my PhD. And my career.
With AI to get a PhD, you can open the door for you, but I warn those pursuing it for money. If someone wants to pay you millions of dollars to work for their company, they will more aggressively screen and filter candidates.
If you can maintain that level of motivation for 5-6 years, it will take you to earn your PhD. However, it is recommended that you apply that motivation elsewhere, as it can be achieved in other areas as well.
Are there any stories you would like to share about your graduate studies and careers with AI? Please contact this reporter ktan@businessinsider.com.
