At the end of the year, I normally conduct a retrospective. This is, however, the first time I’m making it public. Naturally, there are things that I cannot, or am unwilling to share, and hence this represents a subset of my reflection over 2018. Here I share the key milestones and some thoughts in chronological order.

January marked the start of my second semester as a Year 2 Computer Science student in the National University of Singapore. Some time late last year, I took an interest in Machine Learning, working on Carousell’s image search and price suggestion mechanism. I took many ML-related modules, and undertook an undergraduate research project (more on this later). I took a total of 8 modules, spreading myself way too thin. I struggled quite a bit that semester, but managed to pull through because the modules I were taking were interesting and interconnected.

After the semester I took a nice break in eastern Europe, visiting Croatia, Bosnia and Slovenia. Upon my return, I joined Carousell as a Data Scientist Intern, where I internationalized the smart chat replies feature to non-English speaking countries like Hong Kong. I also worked on the pipeline that powers the new image search and price suggestion mechanism. I’m grateful to have been given the opportunity to work on high-impact features, using large datasets consisting of millions of products to train models. Carousell has a great data pipeline that abstracts away a lot of the painful data pre-processing work for me, allowing me to focus on my experiments and getting things to work. On hindsight, I should’ve been more proactive in seeking advice or looking for colleagues to bounce ideas with: that would’ve saved me a significant amount of time.

Before school resumed, I took a short trip to Hangzhou for a Data Visualization Summer School, with a couple of friends. The summer school was poorly planned and had a lot of repeated content. However, we visited a NLP engineer at the Alibaba HQ, and had lots of good food during our stay.

After a tiring first semester, I decided to take fewer modules. These modules were 4k modules, and proved to be quite challenging. During this time, I also continued working part-time for Carousell. This was also the season for internship applications overseas. At the same time, I had to put in extra work on the undergraduate research project I had undertaken at the start of the year, because I chose an open-ended project and went down a research direction where I couldn’t make substantial contributions. Once again, I was spread way too thin, and had too many commitments. Things didn’t go as smoothly in this semester.

I applied to a dozen companies, and the interviews lasted from October into late November. I was not sufficiently prepared for the interviews, and had erratic performance for each interview. This proved detrimental, because the application process for all companies span multiple interviews, and a poor performance in one meant immediate disqualification. Jane Street invited me to Hong Kong for an onsite, and I lacked the confidence and preparation to produce clean and correct answers. What I needed to do better was consistent and intense practice, including both problems on Leetcode and mock interviews. On hindsight, an overseas internship was not particularly high on my priority list, and not getting one is fine with me.

Things that went well

In 2018, I adopted, or was more consistent with a number of practices that I have found to be quite helpful.

First on the list is exercise and meditation. Despite my busy schedules, a morning run always helped start my day right.

Coming in closely in second place would be reading. I’ve read 43 books this year, up from 10 last year. The books I read ranged from textbooks, to both fiction and non-fiction. I read one fiction book and one non-fiction book at a time. I read at night, and this helps me end the day away from the computers.

Things that went badly

This year I’ve been extremely sloppy with deadlines and time management. I started one of my group projects the week it was due, a project for which we were given more than 6 weeks. The situation has slightly improved after I became more rigorous in my Getting Things Done workflow, and I would need to revisit and simplify this workflow.

I’ve also spent many days in procrastination. A large part of me is convinced that this is burnout. I’ve also definitely not slept enough this year, and now know acutely the effects of lack of sleep. I’ve fallen sick more times than I have in the past half a decade. To resolve both these problems, I need to settle into a good day and night routine, and the December holidays is perfect for that.


Here are some of my favourites of 2018.

Books Read

David MacKay: Information Theory, Pattern Recognition and Neural Networks was introduced to me as the companion textbook for my Information Theory module for the first semester of the year, and this quickly became my favourite textbook of all time. Mackay has the ability to present information about difficult concepts (I’m looking at you, Monte Carlo sampling) in an entertaining and digestable manner. This book also draws the link between statistical inference, neural networks and information theory, all three interesting topics stand-alone, but delightful together.

As for fiction, The Three-Body Problem is both thought-provoking for its clever use of scientific ideas, as well as being a depiction of the Chinese culture during the revolution. I also found myself hooked to the Red Rising tetralogy.