 | |
| Image: Murray Neuzerling |
In May, I started my work at ANZ under the AMSI intern program. Until November, I’ll be doing data science-y things in the analytics team in this beautiful Docklands building.
That’s right, I’m making the scary transition from academia to industry.
Except it hasn’t been scary at all.
Sure, working a corporate nine-to-five is a very different experience to the usual t-shirt and jeans academic experience, but there’s been no catastrophic culture shock. Mostly, that’s due to the immensely helpful folks at ANZ who have smoothed the transition for me and two other interns. I cannot overstate just how wonderful these people are.
My work is a good mixture of descriptive statistics and modelling. Without getting into the nitty-gritty, I work closely with the people in marketing and campaigns, picking up experience in that area as I go along. More importantly, I’m picking up experience in exactly how you model stuff in a non-academic environment.
So far, it has been 40% understanding the problem, 40% getting the data in the right format, and 20% actual modelling. The modelling is the easiest part, really. There’s so much reality you need to wade through before you can start fitting models to data. Every single part of the data needs to be understood as more than a number or a string. There are codes that need to be deciphered and parsed into something that a human being can immediately understand. There are quirks in the data around which you need to cautiously tip-toe. There are NULLs. There are so many NULLs.
Before I started here, I hadn’t done much SQL outside of a university. Sure, I’d read through hundreds of pages of theoretical textbooks on relational databases. A relational database is just an algebra, after all, and my entire thesis is on algebras of relations. But I’d had very little practical experience and, in the end, it actually didn’t matter! I picked up more SQL in a week at ANZ than I would have learned in a year at university.
These past few months have given me a chance to reflect on how (un)helpful my education has been. People often tell PhD students that their degree is useless. PhD students understand better than anyone else the perceptions around the use — or lack thereof — of their degree.
I’m in an especially bad position: my thesis is in pure mathematics but I’m pursuing a career in statistics. Yet it doesn’t really matter. The content of my thesis isn’t useful, but the content of almost any thesis wouldn’t be useful here. Thesis topics are so specialised that the chance of any of them being directly applicable to an industry position are near zero.
I can tell you right now that I picked up all the technical skills I needed for this position very quickly. Maybe that’s the PhD at work, but I can’t know for sure. You don’t pay attention to the times you understand something with little effort.
The stuff I struggle with is the stuff that you can’t learn in a university at all. For example, the biggest challenge for me is learning how to communicate with stakeholders. It’s this phenomenal challenge that goes far beyond simply being able to talk to people. You need to synchronise different perspectives so that everyone is on the same page. That means asking lots of questions, and the right questions, too. I don’t have those skills yet, but I’m learning.
After these months, I’m still no closer to working out how helpful my PhD will be in a non-academic career. Are you an academic who successfully made the jump to industry? Are you someone who’s been in a position to hire someone with a PhD before? Tell me about your experience. Are the perceptions about a PhD being useless true to reality?
In any case, the non-technical skills are the ones I need to master soon because I’ve found my industry. I don’t know what happens next, but my time at ANZ is, at the very least, my first step into the world of data science. I’m so grateful to the people at ANZ and to the people at AMSI.
I’ve never felt positive about my future career before, but now I do.
--------------------
Murray Neuzerling is a PhD candidate in mathematics at La Trobe University in Melbourne, Australia. His thesis is titled “Representability and finite representability of algebras of relations.”
This article was first published on Murray's blog: mdneuzerling.com
Murray tweets at: @mdneuzerling
Comments