Do You Need PhD for an AI Career🛣?

4 min readFeb 13, 2023

Today’s post is inspired 💡 by this question on Linkedin:

In a previous post about “How to Tackle Take-Home Data Scientist Interviews?”, I promised to cover more about what’s in the head of a hiring manger.

Today we will focus on one of those topics:

Should you get a PhD?

This doesn’t completely answer the Linkedin question above, but sheds light on part of it. As for “Is self-taught enough (to get a job)?”, we’ll cover that in a later post!

Short Answer

No.

Long Answer

The short answer for my “Long Answer” is: It depends.

It depends on where you are, what you like, and what fits you the best.

Let’s focus on “what you like” today.

What do you like?

If you are interested because of the hype of ChatGPT and the like, or believe that this is the hottest career on the planet, or are looking for high paying jobs, please wait for the next post. 😉

Stay HERE if you are passionate about machine learning, data science or engineering, already in an educational program in these areas, or have just started your career in a quantitative role.

The best way to understand a career is to actually do it, or shadow someone doing it.

This is why a lot of people choose a career similar to that of their parents. ChatGPT explains:

The field of AI is new, so what should you do?

1. Find your role model

This is probably one of the most practical ways to understand the career trajectory of a type of career. Why? Because you can’t perform randomized study of your career to try out different professions. 😭

Here are some examples of successful individuals in the field who can be inspiring role models to all of us.

🚀 Andrej Karpathy — Former Tesla AI Director, founding members of OpenAI, PhD in Computer Science from Stanford.

I recently saw his tweet that he’s going back to OpenAI. You can get the broad stroke view of his career progression from his Linkedin profile. He has been involved in very research-heavy roles. It makes a lot sense to be in those types of roles transitioning from a research-centric educational background.

🚀 Chip Huyen — Founder of Claypot.ai, writer of “Designing Machine Learning Systems”, Master’s of Computer Science from Stanford.

I bought her book and attended several of her talks at various occasions! She’s one of the most inspiring and entrepreneurial minds in the field. She is dedicated to solving industry pain-points for real-time machine learning.

🚀 Jay Alammar — Director and Engineer Fellow at OpenAI competitor Cohere, THE researcher who visualized the Transformer model, B.S Computer Science from the University of Kansas.

His blog posts about cutting edge NLP and Computer Vision techniques are world-famous. I’m not aware of anyone who has beaten his visualizations.

And, 🚀🚀🚀 Radek Osmulski, Zachary Mueller, Hamel Husain, Mehdi Allayari…

The list goes on…

Now, do you think a PhD is a must?

👉 Well, please keep in mind that:

Everyone’s path to success is unique, and it is influenced by many factors including their personal circumstances, skills, talents, experiences, and opportunities.

Success is not a one-size-fits-all concept and it means different things to different people. Rather than trying to emulate someone else’s “success”, it’s better to focus on defining and pursuing your own goals and aspirations, based on your own unique strengths and passions. This way, you can create your own definition of success that is meaningful and fulfilling to you.

Also don’t limit yourself to people who are known in the field. A colleague or a classmate who’s done more than you, can be a great role model to learn from.

👉 Please also keep in mind that:

We can never uncover the full story of a person’s career by simply looking at their Linkedin profile. If you try to do inference on limited data, be careful what conclusions you are drawing.

If more data points are collected, you can potentially perform more scientifically valid studies like the “Career success factors of women engineers in leadership positions” (2020), or the influence of culture (2018), or personality (2001) and family influence (2021).

But I usually find anecdotal evidences good enough to point directions.