When I started grad school, I did not expect to go down the rabbit hole of deep learning.

I wanted to learn machine learning and data science, that was it. There was a good opportunity in front of me, and I took it. I knew I would have to study harder than I ever have, mentally prepped myself to see a lot of Math. But I didn’t expect to be so deep into it that I would start thinking about machine learning in abstracts. I had started reading research papers describing the underlying principles, priors, implicit biases, and so on. I enjoyed it. They were complex, abstract problems and it was so, so fascinating to read about concrete solutions in the form of neural networks and how data flows through them.

Problems required me to read, read and read a lot. Make decisions on which papers are worth my limited time — there are too many, knowledge has become so democratized and such a large push has been given to universities to publish papers. Reading them pushes the boundaries of what I know, tiny amounts of knowledge forming graphs in my head, until that beautiful click that says I’ve finally understood the problem and can begin to think about the solution. And then I go back to reading.

I love it — reading abstracts, going back and forth between papers, finding them circling one another and realizing where one idea stops and where another begins. With every paper I skim, I become more informed, more critical, more confident in my thoughts. I am not an expert, can never claim to be one, and I do not feel the need to be one. I enjoyed the practice of research, as simple as that.

Now that I’ve described my experience as elaborately as that, I now understand why it was painful for me to transition back to the industry. It’s not the tech industry itself nor the idea of it that was particularly painful — in fact, I like it. I’ve always been curious of it, and I’ve been wanting to go back to the “real world.”

But the thinking required to be a worker in the tech industry is vastly different from the one in academia.

The deep, abstract ideas are gone, replaced by actual solutions: practical ones that people actually use and rely on. It wasn’t about researching to find the best solution in terms of the goal, it was about researching to execute the best solution for a fractioned goal: the goal itself, time, money, reliability, scalability, and so on. It is a different beast, one that I wasn’t prepared to tame.

Not that I couldn’t, not that I didn’t want to, and not even that I am clueless on how to go about it. I have an undergraduate in computer science, after all. I know about practicality. I know about software engineering, the tools used in the industry were right outside my current mental vicinity. I read about them in my spare time. But these weren’t enough to prepare me for the context switch, I was too attached to research. The frame with which I previously thought about problems — the one that allowed me to finish my thesis and write two papers about it without any supervision; the one that allowed me to understand what the smartest minds in machine learning in the world are saying — thatmental frame does not fit the industry, and it took time to reshape it. It was painful. Gone were the research papers, replaced by the so-and-so new library or framework that makes it easy to run and deploy machine learning models. No more reinventing or improving the wheel, just use it. What I loved about research had to be pushed back. It was an uncanny state that I cannot fully describe, because it almost made me hate my work, but at the same time I liked my work, was excited to make things happen with what I know.

So, three months in, how am I doing?

I’m OK, much better than my first few weeks at work. I’m learning and using different tools and I’m having fun. I missed this. I missed being a developer, and I’m still as excited to work on different projects and explore methodologies for the real world. I miss deep learning, I still do. I still read some articles here and there on the new ideas, the frontiers, as they would say. Their ideas still intrigue me, and for now I am content with that. My tangible and professional goals anchor me to my work. I don’t feel burdened by it. As I said, it’s a different beast and I’m working on taming it. I’m learning the ropes, each new one a level higher.

Will I go back to graduate school?

I have the capacity to do so, and I only have to reach out to my professors and ask to be referred for a Ph. D. outside the country. I’ve been offered that already. But not now, not yet. I have many reasons to stay where I am, I can work on research projects in my spare time, anyway. So unless I feel the need for a large change of scenery, the answer is No.