Data science student at a wastewater plant doing interpreter

I have no idea how I got here. But they treat me well (in Chinese internship standard), so whatever.

— Loe

---

Intro

Technically, I am not really at a water treatment plant, but I am working at water treatment management company at Beijing. It is a traditional company that has no heavy investment in fashionable technology, like machine learning, or artificial technology. Thus, I am quite unique in workforce repository, a data science student. However, another obvious drawback is that they do not have any related equipment, like reasonably performance-oriented servers, let alone, the fancy GPU ones. Despite these, my boss promised me to get a GPU, and try to push for a proper GPU server later. Sounds great, but both promise are still at horizon as of now. Despite lack of the proper equipment to some extend, I did not really need that much of computation power in the beginning weeks of the internship, because I am starting my internship with basic data cleaning and basic machine learning.

---

Task

I was given a task of analysing a bunch of sensors readings from a water treatment plant from Linyi, Shandong. The sheet has all its features labelled in Chinese with some unclear English letter abbreviation. Super confusing. I does not matter, i thought, I am a data science student. I only need to know the independent variables and dependent variables. “Wait, what am I looking for? “, messaging to the boss. I did not meet my boss in person. He was on the holiday with his family for important family affairs.

After spend some time working with some water treatment professional at the office, I get an abstract picture of what is happening and what the data in the sheet is describing. Basically, the data describes working status of a municipal water treatment. First, sensors at the inlet measure the inlet water flow, chemical oxygen demand (COD), total nitrogen quantity, the total nitrogen quantity presets in ammonia, and total phosphorus. Then during the process, sensors measure the nitrate nitrogen in both treatment line. (Note that one intake line, but two treatment lines, and one outlet line.) After that, the outlet sensors monitor the same matrices as the inlet sensors. Lastly, there are four pump sensors (five, but one is broken) measure the speed of two different types of pumps. First one is the so-called carbon pump, which adds organic carbon into water, removing nitrogen. Then, there are PAC pumps, adding PAC to water, dealing with suspended dirt particles.

In total, I received about 65,000 record, in the span of two years. The general data quality is vert decent. There are very little missing data, and most sensors have relatively short down time. Also, I was told that the treatment plant is a rather well-equipped one, with so many working sensors. Those sensors costs millions in CNY, let alone those additional up-keeping fees to make them function properly in sludge water, which if you ever worked there, you should know it is a very bad working condition.

Despite all my praise to both the plant and the data, the data is still flawed in some manners. Some records still misses certain data, and there are some abnormal readings hidden inside. Additionally, some data like timestamps need to be processed before going into machine learning models. It is too hard for machine to understand timestamp data. Some process needs to be done to the entire dataset. And I was ready to all in, devoting to the task.

---

REAL Task

Then, at that unfortunate Thursday, the unfortunate thing happened. I was asked to give some help in an international conference, as this is what they told me. They said, I am educated abroad, and excel in English. I appreciate their recognition of my ability, but I have only been at the US for a freaking half a year. My English level is only good for Subway, even Panda Express is too hard for me to order something. Anyway, they insists that I go help the professional simultaneous interpreter that they employed. Of course, as an educated person, I was like, wait, simultaneous interpreter is a professional job with high threshold in terms of language understanding and requires plenty real-world practice. Their response was like, whatever, you would be just sitting there helping out. I hesitated a bit more, but I also hinted myself that I was merely an internship worker, they probably would not give me any important job. Eventually, I reluctantly accepted.

I arrived at the conference location. Some time later, the employed professional simultaneous interpreter arrived, in a white shirt, very neat looking. After the not-my-boss boss introduces me to him, the interpreter, I could tell he is more surprised than I am. In the beginning, I could feel his anger, but like five minutes later, after he and the boss realized that there is no way to get a second professional simultaneous interpreter into the conference on time, he just gave up and accepted the fate. Let me provide a bit of context here. The conference is about five to six hour long, and focuses on the water treatment. It also invited a lot of foreigners to give speech, some of whom speak English with different flavours than anyone I meet in the U.S. Additionally, the job of interpreting on the fly is a professional job. People make decent money by doing the task professionally. Moreover, because how entrusting it is to do simultaneous interpretation, the common practice for a long session is to have multiple interpreters working as group, doing shifts in around 20 minutes.

Well, well, well, well.

Then, I was the only choice for him if he would like to, to some extend, finish the task, because he had no way of completing the session alone. It was just way too long. So, here I am, a data science student at a water treatment plant working as a simultaneous interpreter. How cool is this.

Sitting inside the booth at the rear of the huge conference room with a window in front that people can see me, I was, of course, extremely nervous.

Overall, I did a terrible job in the interpretation in the conference, but what a surprise, a first-time newbee doing this. Anyway, I did not care, especially, at that time, because I was so tired, and all I wanted was a feast. To be fair, the professional interpreter did praise me, saying that I did decently well for a first timer. I could tell that he was mostly trying to be nice, but I took that. And, I felt like I did do the job decently for Chinese-to-English part. The English-to-Chines part was not so well. As I later recall, the speeches at the conference were the speeches that I paid the most attention to in my entire life, but I still could not remember a single complete concept from them the second after the conference.

Anyway, I got some iced hawthorn for desert that night. I am still regarding not getting paid for what I contributed to the conference meeting. I mean the minimal payment for a simultaneous interpreter is at least 4000 per day. I complete the task, and I feel I deserve the payment. Well. Anyway. Never mind.