I published this article in Towards Data Science here
Dreams of Research
Three years ago, in a single week, I co-founded the Melbourne University Biological Society and got a job working as a research assistant in a protein engineering lab via CSIRO’s UROP program. I remember feeling like everything was coming together and that I was finally on the path to becoming a researcher.
A 2nd year Bachelor’s of Science student, I was still enacting a plan that I had formulated in late high school. The plan was:
- Complete an undergraduate degree in biochemistry.
- Complete an honours year in a lab with a supportive supervisor.
- Start a PhD and work hard to find insights that might one day save lives.
At the time, I knew a lot of people with that plan. Many of those providing me with advice up until this point thought that someday I would indeed be a scientist. I’d always loved biology, and I was probably smart enough to make it a career, if I worked hard.
Life Sciences’ Research is Awesome…
Working part-time during semester and full time during the holidays, I loved being part of a lab. One week we’d be reading papers and formulating a plan to test a hypothesis and another we’d be problem-solving biochemical workflows.
I’m not sure how much I was contributing, but I was learning lots and I’m grateful to this day for the PhD, honours students, Postdoc and my supervisor for the time they invested in me.
However, it was soon obvious that my skill set wasn’t enough. It seemed like if I wanted to have a secure career, I needed to learn the computational side of life sciences research. This notion was based on discussions about how experimental research couldn’t compete if it wasn’t backed by computational biology or bioinformatics.
With the encouragement of my supervisor, I decided to switch from a biochemistry major to the brand new computational biology major. I’d never hated math and I could see that computational biology greatly depended on it.
It wasn’t long before my brother had convinced me that adding a major in statistics was the best thing that I could do to improve my chances of making it in the world of computational biology, so I took up that as well.
The next year, it started becoming clear that my lab may not exist long enough for me to complete honours. My dream of being a researcher, which had always felt like a north star, was starting to feel like a dangerous gamble.
But the Lack of Funding is Disheartening
You hear about labs losing funding but the reality of what it is like for the remaining labs is something less often discussed.
“Publish or perish” was the mantra I kept hearing. It meant I could expect my future career to involve more time writing grants and trying to get into a good journal by picking popular topics and less time thinking about the core science.
Put all this together and there seemed enough evidence that I should rethink my plans. It was no longer clearer that biosciences was for me. I need to rethink my plans and find a new way to solve interesting problems and make the world a better place concurrently.
A New Dream of Data Science
When the fog finally cleared, I found a new north star and with it, a new plan. With one semester left of my statistics degree and two and a half years after becoming a research assistant, I formulated a plan to become a data scientist.
It had finally occured to me that data science was the avenue by which I could give back to humankind. I could leverage my solid foundation in statistics and experience coding in R and python, to join an interdisciplinary field that was breaking ground in biosciences and beyond.
Looking back, machine learning was already present in many of the projects I had been working on and in the papers that I had been reading.
My plan went something like this:
- Develop skills in Data Science methods and Deep learning
- Demonstrate those skills in a meaningful way
- Get a job working for a better humanity.
I expected to iterate that plan. Learn more about data science, demonstrate those learnings, get a job, learn more, re-skill appropriately. I started with some online courses in data visualisation for data science and in neural networks.
My degree was coming to an end, however, and I had changed my mind way too late in the year to get an internship. I needed to make a decision.
First Steps towards Data Science
I managed to reason my way down to two options, work or keep studying, either formally or informally. I remember having long discussions with friends and family trying to decide whether I should apply for a master’s degree or strike out on my own.
If I went for the masters, I would be implicitly assuming that I couldn’t learn what I needed to online or via a first job. If I didn’t, it would mean spending 3–6 months teaching myself what I needed to know before trying to get a position.
In the end, I decided to choose between Melbourne Business School’s Master’s of Business Analytics (“MBusA”) and Melbourne Universities’ Master of Data Science. The former appealed to me, as it blended familiar statistics with business acumen and data science/analytics, and I was concerned about the latter after speaking to some current students who were concerned that the course was overly technical and a poor proxy for real world problems.
I thought then, and I still do, that algorithms alone don’t create insights and once you have insights, you need to be able to sell them to decision makers. I chose the MBusA to get me thinking about how to use data science in the real world and how to sell those ideas.
Several months, an application and an interview later, I got my offer.
That day my father called me and asked me to reconsider. He believed in me more than I did and wanted me to see what I could do without forking out 55k. I’m still wondering if he was right, because 8 months later, everything fell into place and I dropped out of the Masters.
A Golden Opportunity
In late June of this year, in the midst of an intensely busy and challenging module of the MBusA, my brother messaged me about a data science position at a biotech software startup. I looked into it and the job description immediately grabbed my interest.
I’d been looking at job descriptions all year from companies that wanted everything: SQL, python, statistics, machine learning, 3 or 5 years’ experience and stakeholder management skills for entry level positions. It made me think most companies had no idea what they wanted in an entry level data scientist except that these companies knew they couldn’t teach their recruits and I hated that.
This job advertisement, and this company, appeared different.
The job listed a set of interests like machine learning, graph theory and bioscience as well as attitudes like radical honesty and customer centric design. I’d been doing lots of machine learning recently, had applied graph theory in my computational biology major and was keen to apply all the bioscience that had been accumulating dust in my head.
So I decided to go for it. And I mean GO FOR IT. By the end of the final job interview round, I think I’d written about 20k words. Words I’d written summarising the background of the founders, looking at articles about start-ups, researching product development, summarising old projects or relevant studies so I could sell myself.
Success and a Challenge
In the end, I was offered the position of data scientist at Mass Dynamics.
With 8 weeks of coursework and a similar length internship left on my masters, I’m not sure I even considered finishing. This felt like my chance. In the words of Alexander Hamilton as relayed by Lin-Manuel Miranda — “I am not throwin’ away my shot”.
Two weeks in, being part of Mass Dynamics has met and exceeded my expectations. I love getting to work closely with passionate, talented people.
I feel like I’m learning more in weeks than I learnt in months while studying, but that doesn’t mean I’m not using my previous studies. If anything, I feel like I never know which prior experience, from data science, computational biology or even biochemistry is going to help contribute to the next challenge.
Mass Dynamics is on a mission to accelerate medical breakthroughs by making it easier to derive insights from scientific data. We’re building an ecosystem for the analysis of scientific data that is founded jointly on solid software engineering principles, scientific expertise and human-centred design.
The best part is that despite giving up on my dream of becoming a life scientist I still get to help accelerate breakthroughs that will help humanity which was always the whole point anyway!
But getting over the fact that this new position and my new workplace are awesome, I have also internalised that it is a challenge and a responsibility. Mass Dynamics have given me my shot and I intend to take it.
Advice and Tips
So what is the lesson? What is the takeaway?
- Never stop learning. Never stop asking questions. In the long run, it’s how fast you learn not where you started or where you are right now that will matter.
- Be resilient. Failure is a learning experience.
- It’s not a 0 sum game. Life isn’t a Kaggle competition. Learn from others, share what you can.
- Find valuable mentors. I owe so much, maybe everything, to the people who listened and especially the people who told me to shut up and told me what’s what.
If I could do it all again, would I do anything differently?
Probably not. I often made choices, like learning more mathematics and computing that opened up my options rather than limiting me. I’m not sure if I made any choices that were restrictive and I’m glad I did it that way.
I never doubted it, but this year, it seems especially obvious that humankind needs excellence in biotech to protect ourselves and create a brighter future. Once upon a time, I thought the only way I could do this was a researcher. Now, I think that if you want to help, you can find a way.
Look to those around you, ask what they think and really listen.