What I have learned from One Year in the Industry, Compared to Academics.

Time Lapse of Blue Lights

For most of my life I have been learning, I have a Bachelor and Master of Engineering degrees in Mechanical Engineering, both from universities in Nigeria before moving to Japan to obtain a Ph.D. in System Information Science. After doing a Ph.D., a lot of people would expect that you continue in academics, which is work in a university as a researcher, but I have chosen a different path such as working in the Industry. Now I work as an AI Engineer with a startup in Tokyo, Japan. After almost two years of industry experience, these are the things I have learned so far.

There are two sides to this post. Specifically, I want to share what I have learned so far. I also want to answer the question, do you need a Ph.D. to work in the industry, more importantly as an AI Engineer, Machine Learning Engineer, or Data Scientist?

As a clarification, when I say research, I mostly mean working as an academic in the university. There is another category of researchers who are employed in the industry who are meant to focus most of their time on research and development depending on the company’s vision. These kinds of positions are rear and highly competitive compared to academics in the university. For me, this is like the sweet spot for someone who likes the industry and still wants to continue in research. And this is the kind of thing I do now, which is like a dream come true for me.
The other end of the spectrum are professionals whose main purpose is to help the company to make products. So, you can think of a University professor as opposed to a software engineer or production engineer at company X. These are the kinds of people I am comparing.

I was a research and Ph.D. student for 4 years, a software engineer for a year, and now an AI Engineer for a couple of months. Everything I am talking about here is from my experience and my opinion. If you have a different experience please let me know in the comments, I would love to hear from you.

Industry skills are different from research skills

The industry is about special, time-tested skills and tools, defined by market and production trends, while academics is mostly about contributing and creating knowledge. So as an academic, your work is to chart a new course or find something new, which might or might not later be useful in the industry.

When I was doing my Ph.D., my focus was developing some new methods in my research. Because of this, it did not matter much which tools or programming languages I developed my method in. If it was verifiable and the theory was sound and good enough for publication. As a result of this, I preferred using Python as opposed to C++. Python was easy to learn, fast for prototyping, and trying new things. It meant I could do more research-related stuff without spending a lot of time on learning the language itself. Now compare that with C++. Even though C++ is a more powerful and versatile language, it required a large amount of effort to learn and use. But it turns out that if you are into mass production, you will want to go with C++ rather than Python. One major reason is that it is faster and it is more low level, which is the kind of thing you want as an industry specialist. So even though I had minimal knowledge in C++, when I got the job as a software engineer, I had to upgrade my C++ knowledge because that is what I will need to use most of the time.

Time is a real differentiator

As a researcher, you are more on the creative side, which means you might not get quick results because you are trying to do something different and to do that you need to spend a considerable amount studying, researching, and developing. This is not to say that you can keep working on something forever but you can justify the time you are spending by the result you eventually get. In the industry, you do not have the luxury of time. You need to take the method that works and implement it at once and make a profit with that. This is good for business, and this ensures that you are ahead of the competition. This is what business is about. You do not have the time to develop a new method, especially if you are in a competitive market. Understand that there is a place for research and development, even in the industry but less than 1 percent of companies develop something new, the rest just exploit what has been developed.

Method vs Result

In academics, you want to develop or you are searching for a new method, hence, your result is only a product of this journey. It may be positive or otherwise but you create knowledge in the process. This is valuable as your incomplete ideas can be expanded, continued, or built upon by another researcher. The search for knowledge never ends and monetary gain is not the priority.

In the industry, results are prioritized above method: If you can get a good result with a great method, then that is perfect. But primarily, you need to get satisfactory results. Businesses survive through profits made so the aim is to always get satisfactory results. Therefore, if what you are developing does not contribute directly to making money for your company then it is just a matter of time before you become irrelevant in that company.

That, being said, Research and Industry are like the chicken and the egg. There is a level of research in production but sometimes you need dedicated research to win the competition. Companies who only exploit, that is only focused on production, soon become obsolete or at least become nonrelevant. In most cases, Academics/Research is always years ahead of the industry.

In the case of becoming an expert, you can become an expert through both paths. The hard work and time spent in the industry makes you an expert. You end up doing the same thing over years so this makes more proficient. I have realized also that the number of years on your resume in most cases directly shows your level of experience. And this is where academics is different. In academics, you work precedes you. The years are important but if you invent something significant, this immediately shoots you to expert level. This is not easy and it is rare but it does happen.

So, this partly answers the question about whether you need a Ph.D. to work in the industry. In my case to be an AI Engineer?

You need to think about your goals. It does not make sense to have a Ph.D. and take a job that will not use the research and innovation ability that you have developed over the years.

Do you love the process/challenge more than monetary gains? This is not to say that academics are poor. However, averagely you will earn more in the industry compared to being in academics.

Are you passionate about creating a solution? It is more difficult to do this in the industry unless you create your own company.

What level of proficiency do you want to reach?

There are 4 levels of proficiency that I have found. This from my experience so far in AI but it is certainly applicable to other industries.

1. The Appreciator: He or she understands the conceptual idea, knows the news, knows the history. Knows the trends and will make money from it by investing in the technology. He is the fan and follower, early adopter, and Kickstarter backer. He can also be a critic, connoisseur, or historian.
2. The User: Knows the tools and how to use them. This is a typical worker, software engineer, programmer. He knows how to use the latest tools and he is good at it. This is the guy that does the labor-intensive part, he is mostly in the background and it satisfied with that. He moves the needle and there is no product without him.
3. The consultant: This is like number 2 but with years of experience and a large portfolio, plus not just deep domain knowledge but also wide enough to give expert advice. He or she can organize all of this together and offer it as a service. And it does not have to be professional.
4. The creator: There are two types of creators. The Visionary and the Expert. There is only one thing that differentiates them, Domain ability. The Visionary only has the idea but goes ahead to pursue the idea by employing other people to do the work. He would need to be able to sell this idea to others who then make it their own and help the Visionary to realize it. The Expert on the other hand has both the idea and the practical (or technical) ability to realize the idea. He then starts to work on it by himself. He only gets other people when he has gotten to a stage where he can no longer do it by himself or herself.

With all that I have said, I think most people who work in the industry fall under users and consultants while most Academics are Expert creators or Domain consultants.

I am still learning but this is where I am at in my journey. Let me know what you think in the comments.