Why I chose data science over game dev for my first full-time role
At the beginning of my full-time career, I made a choice between two full time, new grad opportunities - data science as an individual contributor (technical role), or game development in a project/production role.
I’ve been asked about this fork in the road in coffee chats enough times, that I felt it worth writing about. The thought process will be quite specific to my circumstances at the time, but I hope that by writing more of these decision making posts, it can give you some additional perspective in approaching your own career planning.
- Early career preference: Flexibility
- Picking a role that provides multi-category flexibility
- Perusing game development as an independent (indie)
- Comparison, in hindsight
Early career preference: Flexibility
At the time, from my perspective as a new graduate, I knew that I would enjoy gaining a variety of experience in my career.
Apart from polishing what I am good at, I could intentionally continue to expand my skillsets. In other words: “Real experience of 5 years, not 1 year experience repeated 5 years over”.
One can achieve this in several ways over the course of a career:
- Different industries (e.g. Gaming to Finance to Retail)
- Different company types (e.g. large corporation to startup to non-profit)
- Different project types (e.g. ML for website recommendations to ML for robotics)
I encourage you think think about the types of “axes” you might grow from. For example, different geographies may also count - meeting new people and building a new network from scratch improves one’s communication skills, intentional or not.
Picking a role that provides multi-category flexibility
With the above assumption, that in my early career I preferred the flexibility to eventually switch industries or company types, I started analyzing my options.
Not every type of profession can help me achieve this preference. For example, if one chooses to pursue law or accounting, one is locked in for relatively longer.
Economics, which my formal education is in, granted an impressive amount of freedom. I could potentially move between government, finance and banking, even video games (who do you think designs the online marketplaces, lootboxes, and cosmetic skin trading economies?)
This allowed for flexibility between different industries, as well as another category of flexibility, which is between job titles. It seemed possible to also swap between economist, analyst, game designer, research scientist… and so on. Compare this again to law or accounting, which is flexible between industries, since every industry needs lawyers and accountants - but the job titles are much more rigid.
By this point, I was already more set on choosing the data scientist role rather than the project/production role in gaming, which I will elaborate on below.
Data science is transferrable across industries and roles
In data science, job titles can be spread across (data scientist, applied scientist, machine learning engineer, data engineer) and so on, but there is budging room depending how much (Ops, ML algorithms, Software development, etc.) that one prefers, or that the company needs.
In addition, the skills I would learn for ML modelling can be applied across multiple industries. Banks, telecom, ecommerce, and more, all use churn models, fraud models, and so on.
Reinforcement learning can be applied in robotics, ecommerce, and any situation where a feedback loop can be set up well.
Recommender systems are basically in any sort of online marketplace and social media. The possibilities are literally endless.
This is one reason I really like data science as a role - it would not only help me gain high-demand skills, but open even more doors in the future.
Perusing game development as an independent (indie)
Another consideration that I had regarding gaming as a full time job, came from my economics education. Gaming is a cyclical industry - video games are an optional good and therefore likely to be among the first things people cut from their consumption during a tight economy.
Again, as someone making a decision about her first ever full time role, I didn’t want to worry about unnecessary instability while I was building my beginner skills. In hindsight, there is no way that I can prove that it would have hindered my career growth, but I have no regrets that I made the risk averse choice there.
Comparison, in hindsight
Much as I considered dimensions such as flexibility between industries and roles, and macroeconomic factors such as the cyclicality of industries, I do not think that there is any comprehensive blueprint for a 40 year career.
Perhaps in your own career, you have considered these factors and more - and perhaps you didn’t consider the factors I mentioned at all. The criteria is so vastly different from person to person, so I outline my past thought process as a means of documentation and trading notes with my readers.
Looking back now, here is a very gut-feeling assessment of my choices:
Rating for my personal situation at the time. Details in following section.
Full time data scientist
Industry flexibility: ★★★★★
Role flexibility: ★★★★☆
Having fun at work: ★★★★★ - It does vary day to day. For the sake of getting things done, a realistic amount of other types of work is necessary. Even ML, coding, or programming is dull at times. But overall, 5 stars from me.
High control: ★★★☆☆ - This is never going to be as high as if you were working for your very own company, as you’ll see from the Indie developer ratings below. The star count is relative, and higher doesn’t always mean better - it’s just a description of reality.
Weekend indie game developer
Industry flexibility: ☆☆☆☆☆ - The flexibility I can imagine is if I’m making different genres of games. It’s all still under the gaming industry, though. Also, what an “industry” is not black and white - one can make games in EdTech, for example.
Role flexibility: ★★★★★ - Gotta do everything - marketing, business development, errands, programming, PM…
Demand: ★☆☆☆☆ - Perhaps only my own studio demands my work…
Having fun at work: ★★★★★ - Same as data science, this depends on the task and day to day. But overall, 5 stars from me.
High control: ★★★★★ - I call the shots - I can add whatever I want, whenever I want (or get around to it).