Making the Leap of Faith, step by step
Foreword: As I wrote about my journey transiting from an HR Officer to an AI Engineer, I realised what worked for me may not always work for someone else. But I think my story is a little uncommon, so there might be something worth sharing here.
Fun fact: I graduated from NUS with a Social Science degree in Psychology and joined a HR team for my first job. Now, I’m an AI Engineer at AI Singapore.
To some, it seems like a complete 180-degree career switch: from a “soft/artsy” domain such as Psychology and HR, to a “hard science” such as Artificial Intelligence, Computer Science, Math. So, how did I end up here?
When I was a fresh graduate, I thought having a degree meant that I was qualified to change the world for the better. It was finally time to put 4 years of knowledge to good use.
Boy, was I wrong.
In my first job, everyone was always busy. There was simply no time to step back and think about what we were doing. Everyone was just doing things because “it worked”. I even had a colleague who analysed data by printing spreadsheets and reading them row by row with a ruler. So much for living in the 21st century.
There was a strong disconnect between my workplace reality and the hyped up idea of a workplace in the digital age that I had came to believe in.
Being the millennial that I was, I feared the day where I would accept the status quo and live as a mindless office drone. I became frustrated with the situation I found myself in; I wanted to work in a digital workplace, but I couldn’t enable it. I was merely waiting for the digital workplace to happen to me. It was then I realised that to become an enabler, I needed to gain the right skills and knowledge. I started out by looking out for courses, training, and projects to level myself up. I stumbled, experienced failures, but eventually joined the AI Apprenticeship Programme and made a career switch to become an AI Engineer. This awakening was the start of my Data Science journey. Looking back, here’s what I think worked out for me:
1. Work smart by finding entry points into new skills from your current work
Transitions are hard, scary stuff. Not only was I juggling between work and self-learning, I had no coaches or mentors to provide feedback. I wasn’t even sure why I was coding what I coded.
I had to find a way to work smart.
I realised that HR and Data Science need not be mutually exclusive. What if I did both at the same time? I forced myself to integrate what I’ve learned in online courses and apply it to my everyday work.
Prior to my online courses, we normally used Excel spreadsheets to manage data and run analysis. After a few online videos, I made a deliberate decision to use Python for my analysis; Python was recommended as a skill needed by Data Scientists. Needless to say, I struggled with it so much that at multiple points in the project I was sorely tempted return to Excel. Why spend one week googling how to do sums, averages etc. in Python when I just needed a few clicks in Excel?
Obviously I missed my deadline for that particular week, but from then on I performed all my work in Python. I daresay I came out of the project a better programmer who could contribute more to the company.
By forcing myself to apply skills from Massive Open Online Courses (MOOCs) to my work, I managed to invest time to practise my new skills, without sacrificing time or quality of work. While the initial transition can be costly, we need to acknowledge that building skills takes time, and that we would only reap the benefits in the future. In fact, as I applied and practised more, my work became quicker and of greater quality than before.
Of course, this little trick is not limited to HR and administrative work, and Data Science as a skill.
Working in the finance sector? Try analysing stock market trends.
Mechanical Engineer? Try implementing a Computer Vision project for robotics.
I’m just suggesting from my own limited perspective of other fields, but I’m sure you get the point. Be creative and apply the new found skills into your work. You’ll find that, as you get practice in, both your work experience and skills improves.
Work hard, work smart, friends.
2. Be thick-skinned. Your first project will never be good, but it’s important to make the first steps
Around the time when I just finished my first MOOC, my Deputy CEO asked me to develop Machine Learning models to predict attrition within the company. I told him I was still learning, to which he replied, “You’re probably the most qualified in this room, and I’m asking you to try.” 😭
Was I qualified to do that? Probably not.
Did I succeed? Hell, no!
Armed with only 20 hours of training, I wrote about 80 lines of code that achieved ~60% accuracy in predicting who left or stayed in the company (slightly better than a coin flip). I ended up delivering a “machine learning model” that was really a PowerPoint slide hypothesising why the model was bad with no clear ways of advancing the project further. I didn’t know how to improve the model or deliver a product that the managers could use.
Even though I consider that baby project a failure, I learnt crucial and practical lessons that I didn’t get from a classroom setting.
You’ll strengthen your understanding only by trying out things and applying knowledge. What worked? What didn’t? Most importantly, what were the gaps you identified? The last question is especially important because it helps you to chart your next steps for improvement.
Adult learning is all about learning through practice (see: experiential learning). Be thick-skinned and accept projects, even if sometimes you feel like you are not ready for it. Your first product will never be good, but you’ll make iterative improvements to the project, to yourself. And that’s really the point; it’s all about trying, taking the first step, and letting the next few steps unfold with clarity. If you’re not already in the industry, this is the closest thing you can get to hands-on experience that everyone talks about in an interview. People want you to solve their problems, not impress them with textbook knowledge. So it’s important to get familiar with some of these actual problems they face. Get your hands dirty.
Of course, you need to be responsible and communicate to your stakeholders that you are still in the learning journey. Some bosses will probably still agree to letting you take on non-critical projects, but it’s good to be honest from the get-go.
3. Be your own coach. Let your experiences guide you as you transit
Starting out, I felt alone in my journey. My peers weren’t that much interested in it, and reaching out to established communities was daunting (e.g. Data Science SG, Girls in Tech Singapore). When practising on textbook examples, I wasn’t sure what I was doing; there was no context, no relatable goals, no problems to solve.
It was when I started forcing Python into my work that I began to realise I could try to solve HR problems with Data Science. By doing so, I could cross-reference my Data Science outputs with my knowledge from HR. The magic of learning a new skill comfortably in your own domain is that you can always refer to prior knowledge. This will be helpful in checking to see if the application of your new skill is correct.
Here’s an example: Let’s say I want to investigate how my salary compares to the rest of the organisation. I would use Python to load the data, calculate the average salary of my peers or officers in my department. Whenever my code churned out a result, I could always fact check with what I already knew. In this case, I know that the average salary was $X amount, so if my code returned $Y, I would know immediately if there was something wrong. This feedback was useful in getting me to relook at my code, find the errors, and learn from my mistakes.
Through this back-and-forth process of implementing and fact-checking with my prior knowledge, I had become my own coach and was able to build confidence in my new skills. As I advanced, I could ask deeper, more difficult questions, and eventually my skills also evolved to meet the (self-imposed) demands. My HR persona had become a guide for my Data Science self. I had started my transition into the Data Science domain by bringing my “HR -self” along in my journey.
Some expect change to be a sudden, Cinderella-esque moment when you wake up and suddenly you are living the dream. Reality and change itself are hardly that magical.
Where I think the magic really lies is in the constant upgrading of the self. Think of it as though you are playing a video game: you don’t really know if there is treasure in the dungeon, nor the dangers that lie ahead. What you could do instead to increase your success is to prepare yourself, get the right skills for the job. I believe that when you build upon yourself, you will be ready when the opportunity appears.
With that, I leave you with a quote from Neil Gaiman, and hopefully you will soon take your first step:
I hope that in this year to come, you make mistakes.
Because if you are making mistakes, then you are making new things, trying new things, learning, living, pushing yourself, changing yourself, changing your world. You’re doing things you’ve never done before, and more importantly, you’re Doing Something.
So that’s my wish for you, and all of us, and my wish for myself. Make New Mistakes. Make glorious, amazing mistakes. Make mistakes nobody’s ever made before. Don’t freeze, don’t stop, don’t worry that it isn’t good enough, or it isn’t perfect, whatever it is: art, or love, or work or family or life.
Whatever it is you’re scared of doing, Do it.
Make your mistakes, next year and forever.Neil Gaiman, My New Year Wish, 2011
Top image : Celebrating Chinese New Year 2019 at AI Singapore