I haven’t blogged for a long while – life has been busy, but truth be told, I miss it. As I was going through all my unfinished articles in the Draft folder, I realized in order to write and publish more, I needed to remember what Winston Churchill once said, “Perfection is the enemy of progress.” I need not to worry about perfecting all my articles – just sit down to reflect & write. With this realization, I’m publishing my unpublished articles with minimum updates.
In July 2018, I had the great pleasure of joining a 3-month-old data science team at Cigna on a mission to use machine learning to transform underwriting (The Road Less Traveled: My Decision to Explore Data Science). For two and half years, working alongside data scientists and engineers, I built some cool predictive models using billions of data points. It was a rewarding experience! While in the data scientist role, I wrote down 4 aspects of data science that I greatly appreciated.
1) Abundant freedom to think differently and challenge the status quo: Data science is much more than number crunching using fancy techniques–it’s about looking at problems from unique angles and fundamentally do things better & faster. A curious mind who loves innovation, I appreciate my creative data science work.
2) Life-long learning mentality and endless opportunities to learn & use new technologies: Since taking on the data science role, I have embarked on a fun educational journey. From studying machine learning, Python programming, reading university research papers, to using cool tools such as Hadoop and DataRobot, I realized the more I know, the more I know I don’t know. I feel like a sponge and eager to absorb as much knowledge and as quickly as possible.
After realizing our existing Python programs were too slow and unstable to handle large data sets (we frequently handle data with billions of rows and hundreds of columns), I researched, prototyped, and presented a PySpark-based solution to the data science team streamline big data processing (parallel processing is amazing!). It was very fulfilling to not only dramatically cut down our program run time, but also empower my teammates by adding this new new programing language to their toolbox.
This intrinsic desire to continuously learn & grow is so prevalent among the people I met in the data science (and the broader AI) space. It’s quite inspiring to learn that Andrew Ng, the frontrunner of AI who led Google Brain and Baidu’s AI team, always has a folder of research papers in his backpack so he reads whenever he has time.
3) A sense of mission: knowing the future role of AI in our society–its potential benefits and threats, I often conduct thought experiments and ask myself what would Elon Musk do if he were in my shoes right now. I feel I’m on a journey to fulfill a humanitarian mission (not sure what it is yet, but I’m excited about it). Elon Musk has a highly controversial philosophy around work ethics, but I can totally relate when he said working 80-hour weeks is what it takes to change the world. When you truly enjoy what you do and see the big “why” behind it, “work-life balance” is no longer a relevant concept. Work and life are not meant to oppose each other; work should be a fun & fulfilling part of life.
4) Highly collaborative: in data science, you not only work closely with other their data scientists, but also partners in IT and business.
- IT partners help us to do Extraction Transform Load
- Business partners help us define the business challenges and set up requirements.