How I did my job interviews (Bear Blog Carnival)
Many aspects of the human life has become so much more awkward in the modern age. The most obvious is the courtship process, which may put the peacock to same nowadays. But today I want to talk about the humiliation that most (not all) students have to face: seeking employment. As my former roommates could all attest, I am kind of an impossible person. But somehow I managed to seek the exact kind of employment that I need without having to send out hundreds of resumes like my peers. I won't pretend that I am some kind of guru, because frankly speaking, I still don't really understand how I got what I got. But these incidents of mine may be a not-so-lousy read for this Carnival.
AI startup no. 1
I believe it was the first semester of my third year. I was looking for internship to fill the compulsory requirement for graduation next semester. At the time, I was studying chemical engineering, but I didn't really like my major, mostly because of the prospect of working 12-hour shifts in manufacturing plant like 80% of my peers (who managed to get a job). So as everyone else did, I try to go into machine learning instead.
I found this startup working on computer vision. At that time, I was studying on my own. I did not attend any hackathon nor had any previous professional experience. I did not have anything to show, and did not know how I would convince them through the Zoom call. But I remember being there with the CEO of the Singaporean office, the HR and the head of technology. Whatever nonsense I spouted, it was impressive enough for them to invite me for an in-office programming test. They wanted me to solve a LeetCode dynamic programming question live. I actually got 99% of the tests to pass. The last case is because the optimal dynamic programming will give you a linear time solution while I was doing some unoptimized quadratic time version, and I did not know enough to optimize it.
They liked me enough despite my blatant inexperience. So they gave me a project-based contract. Long story short, I royally screwed up, missing the first fundamental point of Machine Learning Engineering (capitalized 😭). In the end, they decided to end it early and gave me $200 out of the promised $1000. So the first job I ever got, I failed it, but not too badly, I guess1?
Green energy startup no. 2
I could not say the same thing about this one. After the period with the startup above, I still need to look for internship for the next semester. Besides machine learning, I was hugely into startup at the time. I found this startup working on green energy in the country where I was studying. It was actually pretty funny how I got to know them. They got listed on a startup database in Asia. Then I saw them again at the startup event in my school. I also met someone who closely worked with them in the past. I thought that it was some kind of "destiny". Jesus Christ. The interview happened in 5 minutes. I was chased out of the Zoom call. Later, the founder sent me a long email (I asked for it). Basically telling me to get my shit together and know what I want to do so as not to waste other people time2. In the end, I went to the next place I interviewed at. And the analyst position at the startup was filled by a business student from a different university. Good for him.
Gov agency no. 1
This was my 15th application during the internship period. It was for a data engineer position at a government agency. 14 places (excluding the above 2) either ghosted me or returned the most generic rejection email possible. This last place conducted interview via assignment for data engineer internship position. They wanted me to do some simple data explorations and build a data pipeline using Apache Spark. I did not know this framework at the time, and I quickly figured out that this would be a pain in the ass to get an environment to test my code since the backend is Java(!) However, I discovered that they had Python APIs, and I just learned Docker at the time. I put two and two together and found a docker environment for Spark with Jupyter Notebook exposed to the user. The next problem was I did not know anything about the Python APIs of Apache Spark. However, ChatGPT already matured at the time (compared to when it was first released). Once I realized that I could get coherent code from it, I proceeded to a loop of copy-paste-run-repeat and finished the assignment in 6 hours just before the deadline3. As the finishing touch, I declared openly that I used ChatGPT to complete the assignment.
They loved it. After a procedural interview, I got hired to work the next month4.
Semicon Big Corp no. 1
They are the biggest employee for my major. But I'm getting ahead of myself.
After the internship, it was September, and the final year of undergraduate. We had two semesters left before we would be kicked out of the school. We had to file a job before graduation. Now, I am a foreign student so I had an additional obligation to find a full-time employment in Singapore for at least three years. So unemployment was not a choice for me5.
I don't want to work for any conventional employers of my major. I don't want to do shift work in a manufacturing plant in the middle of nowhere. However, again, unemployment is not a choice. I was among the most technical sound students of my batch. And these conventional employers, they hire mostly based on academic results. So, in theory, I just need to apply.
In reality, it was post-COVID adjustment period. The recruitment process is much much more stringent. I decided I need professional help. So I consulted the school career coach service. I really like my coach. She was friendly and helped me flesh out my interview answers and resume which was very highly unconventional for the type of job I am applying to. I actually got to the final round where I directly met the director of my plant, also the director, my senior working with. We had a friendly chat, kind of like the and the procedural interview with my last internship. I thought that I got hired and started to prepare myself mentally for the day ahead.
However, right after that day, a company-wise hiring freeze happened, all existing recruitments not yet closed were dropped immediately. That included mine. To say I was plunged into despair is an understatement. However, life indeed had another plan.
Gov agency no. 2
In the final year, students have to sign up for some compulsory electives to meet the requirement for graduation. I could not imagine myself enjoying intellectual property, pharmaceuticals, or oil and gas. So I signed up for a new class in data analytics taught by a new professor just transferred to my school that year. In a typical university student fashion, I did not attend any of the lectures and speed-watched them 1 week before the final exam. But during the study, I realized that I really like this new professor. I dug up his background and found out that he got his PhD from MIT. He transferred into the chemical engineering department, but his focus is an integration of artificial intelligence and microscopy. So I struck up a conversation with him after the exam. In my typical fashion, I started it by stating the exact fact above: that I like his lecturing style even though I did not attend any of it blah blah. In the spur of the moment, I asked if he was looking for PhD students.
He invited me to discuss more in his office afterwards. The conversation somewhat resembled my interview with the green energy startup. I had zero research experience and my intention of pursuing research as a career only came from the fact that I don't want to be unemployed6. But he was nicer than the startup founder7. That was December.
In May, after the disastrous attempt at the semiconductor big corp, he contacted me inquiring whether I still want to pursue a PhD. Our department actually got some funding and he recognized my enthusiasm despite my inexperience. So I started volunteering at his lab to build my profile for the application the next January (it was June then). He also told me to apply to the government research agency that he was working at before transferring to my school. Frankly speaking, I did not have much hope since there were just a few openings.
But, yes, but, one of the opening came from a project that was struggling to hire. They need someone knowing both programming and chemistry. That was why they struggled to hire, but that was also why it was the most suitable position for me. I had to get this job. I had to get this job. I had to get this job.
What I did next changed everything. I asked about the format of the interview and received the vague answer "we will discuss things". I remember the last successful interview I had, which was for the internship. It was an assignment-based interview. So chances were, to succeed at this interview, I had to turn it into an assignment-based interview. What is the assignment? Well, I was hired for a research project. So the implicit assignment is I need to know what is that problem, the potential challenges they are facing, as well as possible solutions to address that. In technical terms, I need to conduct a background search on my supervisor and a literature review on the problem we would be working on8 and propose potential solutions. I came to the meeting with a slide prepared for introduction, the key problem, the likely challenges they are facing, together with potential solution. The slide could be presented in 15 minutes with 5 minutes questioning afterwards. But after my presentation, I knew that I got hired already9.
A laughably optimistic timeline prepared by a person with zero research experience. Alas, I did put my name in the manuscript we submitted 4-5 months after my start date.
That was how I got my current full-time job.
Ending
My experience aligned with this Tweet. The time I succeeded were the time I proved definitively my worth to the employer. One time the chance was built into the process, the other I had to create it. And with much more luck than I can comfortably admit. If I have to repeat this in 10,000 parallel universes, I don't think I can get a in at least 100 of them. That's something to improve for the future. Right now I am gonna sink my teeth into my job and never give you up let it go, for the next 3 years at least.
Footnotes
I was still active on LinkedIn at the time. I wrote about the experience then.↩
I could not bother digging up half of my inbox in Outlook for this particular email. I assured you that the email was much more formal and polite than what I wrote here.↩
They gave me 72 hours, but I spent the last few days preparing for an already-doomed LeetCode assignment with TikTok since, well, it was TikTok.↩
The official internship period started in January but I was starting in December. I already liked the place and I hoped to convert it into a full-time offer after the internship. Alas, this plan never came to fruition.↩
At least, I did not seriously consider it until April/May the next year, where I was constantly being called to the dark side every day.↩
In case you haven't realized it, there is a saying in Vietnam to describe people like me, which can be roughly translated to English as "Say things that most people left unsaid". I learned of it through a short story about a lone meteorologist working on a mountain and pick individual of nowhere and the conversations he had with the sole driver that drove the few tourists to the region and the few journalists that came to interview him. Whether my trait is a good or a bad thing highly depends on the occasion. Personally, I found honesty is indeed the best policy on average, though I have become more "strategic" starting my full-time employment since getting fired is not an option for me.↩
I later found out that he experienced somewhat the same thing as me. His background was in electrical and biomedical, but his first PI accepted him into his laser lab though he had no experience with laser, and his second PI accepted him into his AI lab though he had no experience with AI.↩
In colloquial terms, I need to stalk my supervisors' Google Scholar pages.↩
On the first page I wrote that I did not know any computational biology but I could learn very fast. The rest of the presentation (I hope) proved that since I produced it in 5 days.↩