The 2018 thriller
In December 2017, we gathered with a bunch of college friends in Dolomites on a skiing trip to celebrate the end of a year together. The trip combined a few valuable things including physical exercise, being close to the nature, and catching up with friends. The joy of daily skiing or solving blockchain intellectual puzzles seemed like a perfect way of entering the New Year. Yet this idyllic scenery was just the calm before the real storm.
Looking back, the year of 2018 was, in fact, like a roller coaster ride with many sudden drops and tight turns. It challenged me in a number of ways both personally and professionally. I made a few terrible mistakes yet I saw the light at the end of the tunnel when I experienced growing pains for the first time in a long while. In this post I would like to reflect on a few things that happened in 2018, and discuss what I am hoping to improve on in 2019.
Becoming a tech lead
This time last year I was wondering what my project focus for the half should be given that I had already spent a year on the team and had acquired a lot of context (at Facebook people usually spend 2 years on each team). One of the ideas that was on the table was building an auxiliary bootstrap model for our main ML system to improve performance in a targeted manner for our clients. At first I was a little bit reluctant as a similar effort was tried to no avail a few months before and eventually stalled. Intrigued, however, with the prospect of building an end-to-end ML solution complementary to the existing product I took up the challenge. My system knowledge and context helped me significantly with execution since I had to touch our whole stack from client-side to backend C++ service and all the way to Caffe2.
Over the course of the following few months we had our ups and downs in the project but during those challenging weeks I realized what it was that made me go to work every day. It was the sense of ownership and a technical challenge of building a proof of concept. The solution, if proven successful, was going to unlock a downstream product for new customers which gave me extra motivation. In the end we managed to successfully finish the test and decided to double down on the project in H2 with me tech-leading it. All of a sudden the ability to enable others became more important than my own execution. Most of all I had to learn to give away ownership and distribute control.
Yet not really
When everything goes up and to the right, it is hard to find any learning signal. As we were making progress with plenty of time ahead, we were on track for the goal. Nevertheless in November, we hit a series of roadblocks external to us. At Facebook, the end of the year is usually stressful due to close proximity of multiple holidays carrying higher user load. For all the reasons combined, we ended up in a very tight position for the end-of-year December sprint.
In retrospect I wish the “when the going gets tougher, the tough get going” saying applied here. Yet for the very first time in a very long while I felt that the imposter syndrome was real. For the first time I hit diminishing returns of spending extra time on the problem as I became more dependent on others. With each move I was worried I was overstepping too much, micromanaging, or getting too involved. Having been used to be able to influence projects I was responsible for directly before, I was powerless. Every very time something went wrong I felt I was failing myself and others while being unable to fix things by pure means of my own keyboard.
Seeing light in the tunnel
One always hears those horror stories about bad managers who may have very negative impact on one’s work and career. Thankfully I have been pleased with my experience at Facebook so far as I have been challenged constantly and received continuous support for me and my growth. I learnt it first hand how much it means to have someone who keeps believing in and pushing you.
On a Friday afternoon just before Christmas, I had a longer than usual 1:1 with my manager over a VC. We had an enlightening conversation when he shared his experience of being a tech lead. It not only calmed me down, but also gave me a couple of ideas of how to improve going forward. In particular, he pointed out the need for better project transparency when it comes to collaborating with other teams. Most importantly, however, he stressed that assessing, handling, and communicating risk is of utmost importance to a tech lead. Research projects may fail since they always carry a degree of uncertainty but everyone dependent on them needs to be aware of their status and they need to result in learning. After all, arriving at a negative conclusion is a positive outcome if it informs further research direction.
Going into 2019, I am looking to improve on tech leading a project based on learnings from the first whirlwind half in the role. I also need to tackle increasingly more complex technical problems myself to keep gaining applied Machine Learning experience. As always I would love to hear more about others’ experience and bounce back ideas on similar problems. In fact, my former coworker has already added one item to my reading list for this year.
Paying back old debts
It is inevitable that each software project accumulates a so-called technical debt. Over time it gets increasingly more expensive to pay back through fixing old designs and code refactors. Nevertheless social and people problems are usually harder to handle for the complex emotions involved. For a while I have been too weak and foolish to tackle a complicated situation that I have gotten myself into. Looking back I know I hurt a few people and I deeply regret it. This year, however, it was eventually resolved forcing me to start from the clean slate.
Painful as it was, it made me reorganize my life and rethink how I was spending it. I noticed that I no longer enjoyed things that I used to such as sports and discovering new places. Even though I was excited to plan and organize a Thanksgiving trip to Shanghai, when it came to boarding the plane I could not help but wonder what was I doing there. I was lost trying to do things that made me happy before. Or that I thought used to make me happy?
My immediate reaction to the ambivalent state was to read more as I craved external intellectual stimuli. I finally found time to look for answers to questions posed when hiking around Machu Picchu. It turned out I was a little late to the game to discover Guns, Germs, and Steel and Why Nations Fail and learn more about development of civilizations. I also read a book recommended by my coworker which altered the way I look at how people discuss things with each other. Only after I read Crucial Conversations I realized how often one communicates inefficiently. Having read it I cannot help but notice how bad I can be in handling discussions even with my immediate family. Even though some of my non-engineer friends did not agree with its usefulness, I highly recommend taking at least a look at it.
Changing society encouraged habits
Having to rediscover myself again and look deeper inside, it was high time for more radical changes. One Saturday afternoon when jogging across the Williamsburg bridge, I decided to try a new social experiment of not drinking any alcohol till the end of the year. It is not that I used to drink a lot before, on the contrary I barely drank a beer or two when catching up with friends. It was more about trying to identify my own limits of withstanding the social pressure and seeing life from the other side.
Interestingly enough, I started noticing how much of one’s social activities in New York involves drinking. As I started seeking things beyond that I felt deep down inside the need for more meaningful conversations with people. I would still go to speakeasies for live music even if it meant ordering non-alcoholic drinks while awaiting waiter’s judgmental facial expression. Most of all, I regained appreciation for simpler but priceless pleasures such as having one-on-one walks along the waterfront.
Giving up alcohol entirely let me also realize how much it can affect one’s sleep quality. Thinking about the adult life, there is overall not much spare time outside work. The ability to claim back control of one’s weekends makes a huge difference as it comprises the majority of spare time over the week. On multiple occasions I surprised myself how I was able to get up early for some physical exercise be it tennis or cycling in the morning.
Winds of change
Visiting my brother in San Francisco last week, we were joking that we can’t live too close to each other. Knowing that he is going to move to Boston in a year or so, I would have to switch coasts again. Even though the year of 2019 seems to be bringing a lot of change I don’t think they would be as radical as moving back to SF just yet. Starting off a new apartment search with roommates, I am planning to strengthen social bonds by spending more meaningful time with others. As such I want to continue exploring natural wonders of Americas since hiking together is one of the best intellectual and social stimulants.
Having been promoted last summer to a Senior Engineer role, I reached the goal I set for myself a while back. For a few days I felt as if I lost all of my energy and had to fill the vacuum of my personal goals space. Yet, shortly afterwards, I found that unsettling feeling of being uncomfortable at work again as I broadened the scope of responsibilities. I am looking forward to overcoming the imposter syndrome and being more effective in the new role while helping others grow.
