My Journey from Econ PhD to Tech — Part 1: Interview prep + Networking

Scarlet Chen
11 min readDec 14, 2020

In Oct 2020, I wrote a series on ‘how to break into Tech’ for Econ PhDs. That series was written with a very objective tone, meant to be a guidebook. However, I’ve always found reading other people’s stories very useful. For example, it wasn’t until I read Haseeb Qureshi’s ‘From AppAcademy to Airbnb’ that I realized how the job hunting process in tech actually looks like. So I decided to share my own story of ‘From Econ to Tech’ to help make yours easier.

Links to other parts:

Part 2: Interview experience

Part 3: Wage negotiations part 1

Part 4: Wage negotiations part 2 + Decision

Tell us about yourself

I’m a PhD in Economics at Stanford. In Jan 2021 I’ll be joining Google as a Data Scientist, working on bidding algorithms under Display Ads.

Why not academia? Why tech?

There has always been two things I wanted to do in my life: (1) understanding how the world works and (2) making positive changes in the world.

During my undergrad years, I spent a significant fraction of time building the HKUST Chapter of CEN (China Entrepreneur Network), an organization focused on using social entrepreneurship (i.e. firms whose goal is not to maximize monetary profit but social impact) to solve social problems. — that is 2.5 good years that I spent on theme (2).

The PhD in Economics is 4 good years that I spent on theme (1) — understanding how human society works. Economics is a social science, i.e. a science. Science’s goal is to understand how the world works, not to change them. I had a great time thinking about how society works, but it also made me realize how much I care about solving problems in addition to thinking about them.

Why do I think tech can help me achieve this goal? It’s less about tech, more about being in/running a business in general. My eventual goal is to address people’s problems by doing my own start-up (more on this later). So the question is, how can I best utilize my current skill set/credential, i.e. econ phd, to get to where I want, i.e. a start-up? Well, becoming a data scientist in tech — the econ grad school training is precisely what the data scientist job needs, and the tech industry knowledge prepares me for my start-up career.

Why not change the world by going into policy/politics? If I were in China, I might think about doing that, because it’s a place where government has real influence on the way people live their lives. In the US, my impression is that the checks and balances also made it hard to induce changes. As a result, it’s surprisingly hard to get things done through politics/policy.

How did you tell your advisors about it?

I had a very easy time because all my advisors/committee were cool.

At that time (end of my 4th year) I was meeting with my advisor once every 2 weeks on my dissertation, and before every meeting I would send him a few slides on what I wanted to talk about.

After I decided to go into tech, I prepared slides for our bi-weekly meeting as usual, just that this time not about my dissertation but about why I decided to go into tech.

After I sent the email, he very quickly replied saying although he is sad to see me leave, he 100% supports my decision, and said that given the situation, I should just graduate now (i.e. that summer quarter) since I have enough papers (3 papers, to be precise), and then I can focus on getting a job.

How did you get started collecting information?

Although econ grad school prepares us for a data scientist career in tech, to actually get a job, there’s still a lot one needs to learn — not just skills like python and SQL, but also understanding the tech world in general — how is a tech firm set up, what’s the universe of job (or job titles) a econ phd can reasonably apply for, what’s the differences between them. When breaking into a new field, the hardest part is that you don’t know what you don’t know.

To be fair, it wasn’t particularly hard for me because RE, a previous Stanford econ phd, organized an ‘economist in tech’ panel in fall 2019 (i.e. in my early 4th year) where 5 economist/data scientist from Facebook, Amazon, Uber, Snapchat, Thumbtack talked about most of the questions an econ phd interested in tech would want to know.

So as soon as I decided to make the jump, the first thing I did was to contact RE — asking her for resources for the interview prep, and the contacts of the panelists, which she gladly shared, and became an important starting point for both my interview prep and networking.

Historically, every year after the job market, the department would hold a job market panel where current candidates share their experience with the next generation. Starting from my year, the department started to hold the academic one and the tech one separately, due to the large number of people interested in tech.

From that panel, I not only got to know who went into tech last year (which I later approached for referrals), but also got to know how the job hunting process look like — for example, I did not know about NABETEC or Kaggle or SQL, or how important networking is, or that you can skip the AEA and look for jobs by yourself.

After these 2 events, I’ve solved the don’t know what you don’t know problem mostly. I know that:

  • If I seriously want to go into tech, seriously preparing for the interview is important, because tech interviews ask for things like python, ML, and SQL, which I know nothing about
  • Networking and referrals are important

Now it’s just a problem of learning those skills and getting those referrals.

How did you prep for the interviews?

I’m someone who prefers to learn things in a systematic way, i.e. through a course. But at that time, I clearly don’t have time to take academic courses from Stanford or Coursera to learn about Python, or different types of ML, or algorithm and data structures, or SQL. In addition, courses taught by universities tend to be too theoretical, not practical enough, and contains too much details. If you’re interested in learning something just to find a job, there are much more focused materials: Data Science bootcamps.

At the beginning of June, just as I started thinking about going into tech rather than academia, I came across a WeChat article marketing Laioffer, a start-up founded by a group of Chinese FAANG engineers/data scientists that teaches you how to find a job in tech. They were recruiting their summer cohort for the ‘Break Into Data Science Bootcamp’, which costs $6300, and takes 2 hours everyday for 5 days a week for a total of 3.5 months.

Now that I think about it, it’s clearly way too expensive. But at that time, I thought, ‘each class is only $50, which is cheaper than my personal trainer’. — what I forgot is that the personal trainer charges $60 for 50 min to train just me, but the Laioffer class is a 1-to-200 zoom session with no personalization whatsoever.

The Laioffer class turned out to be a disappointment: the material was way too basic — it’s designed in a way that even if you have never heard of mean and variances you can still follow, and way too much of the content was on programing — 40% of the class time is devoted to algorithms and data structures, which probably accounted for 5% of the interviews (according to my own experience).

However, taking that class did give me peace of mind — I got to know the universe of things that could come up in a data science interview. Once you have a glossary of things that you should know, actually googling around to learn it is easy. Another thing that I thank the Laioffer class for is the change in mindset — from ‘I’m an econ phd trying to break into tech’ to ‘I’m looking to break into tech’. When trying to make a transition, the destination is more important than the origin.

One other thing I was annoyed by the Laioffer class about was how little hands on experience I was getting in ML — they marketed that students will have 3 ‘hands-on’ project experience, but it turns out it’s just the teacher walking through a Google colab notebook during the class.

I started googling around on my own to look for other bootcamps that are more focused on ML and project experiences, and found Elite Data Science Academy (EDS). I enrolled in their ‘Machine Learning Accelerator’ and got their ‘Interview Prep Kit’ for a total of ~$400. Till today I think it’s one of the best courses I’ve taken during my grad school (the only thing that is a match is Monika and Martin’s part of the first year macro sequence), for how practical things were.

Before EDS, my knowledge in ML was non-existent — I thought ‘unsupervised learning’ means that when you run it, you don’t need to sit in front of the computer to supervise it. After EDS, I not only have 3 ML projects I’ve done from start to finish that I’ve put on my CV, my level of ML knowledge was good enough to get me a job titled ‘Applied Scientist, Machine Learning’ from Zillow’s iBuying business Zillow Offers. (Well, to be fair, they hired me for my econ phd and domain knowledge in housing; but still, I wouldn’t have passed their ML interviews without EDS.)

To summarize, here’s how I prepared for different parts of the interview:

  • Algorithms and data structure: Laioffer class+using their internal tool ‘Laicode’ (very similar to Leetcode)
  • SQL: Laioffer class + Laicode + Leetcode + w3school
  • Python — Data manipulation and numerical computation: Laioffer + EDS + googling around + doing Quora’s data challenge in python instead of Stata
  • ML knowledge, and ML projects to put on CV: EDS class+googling around
  • Experimental design/causal inference/Stats: Laioffer class + googling around + reviewing Guido’s second year sequence notes (which turns out to be not super useful because it’s still too theoretical)
  • Behavioral questions: googling around + thinking hard about my past project experience
  • Open-ended case study: watching youtube videos of talks that the firm’s DS team gave in universities (the recruiter sent me links; and on youtube, once you start watching the first video, it gives you recs for similar videos); reading the firm’s data science blogs/papers; talking with Anthony for 1 hour about the firm’s business model/what questions they might be solving

How did you get referrals?

  • I pinged the Stanford econ (or GSB) phds who were just done with the market and went into tech (whom I either already knew or got to know through the department’s ‘job market candidate panel’).
  • I looked through Stanford econ department’s past placements*, identified everyone who went into tech, searched for their email online, or LinkedIn direct-messaged them. Some replied; most didn’t. (*note: placement = where an econ phd end up in her/his job search)
  • I contacted the panelists in RE’s event whom I thought were not too senior.* (*note: Now that I think about it, I really should have contacted all of them. In academia if you send a cold email to a professor at a good place asking for a 15 min chat to learn about ‘career in econ’ I highly doubt anyone would reply. In tech, in general, things are less hierarchical and people are more friendly/open minded.)
  • One of my advisors put me in touch with one of his former student who is now in tech.
  • During this period of time, I posted some questions on WeChat 朋友圈 (similart to twitter ‘tweets’) about job hunting (e.g. “what on earth is a ‘applied statistics case study question’ in a data science interview???”), and some of my friends, upon seeing them, private messaged me to offer help and/or ask if I’m interested in applying for their firms.
  • Anthony posted on his Twitter ‘My wife who is a econ phd at Stanford is looking for data science jobs in tech. Ping me if you have openings or know who’s hiring.’ and got ~10 replies.*

*note: at that time he was super excited and was like ‘look at how many jobs I’ve found you!’ But in fact, none of those eventually converted into an interview — it turns out that strong VS weak connection makes a difference. For example:

One of the replies was from a data scientist at Doordash — he replied Anthony’s tweet with a link to a job posting: a Senior Data Scientist position at Doordash. Anthony private messaged him and got his email. I then sent him an email with my CV, asking for a 15 min chat, which was never replied to. A few days later, I received a rejection email for the ‘Senior Data Scientist’ he posted— he directly referred me to that position! It’s a very sad story because only if we’ve talked, he would have known that the relevant position for me is the new grad data scientist position (typically, any position with a ‘senior’ in its title requires some amount of full-time work experience, unless you were a faculty before you jumped into tech).

On the other hand, when this happened with a friend of mine that I’ve known for a long time, things were different:

HG, upon seeing my WeChat posts about data scientist interviews on Aug 26, asked me if I’m interested in applying for the firm he’s at — Robinhood. I sent him my CV, and a few days later, received a rejection for their ‘Senior Data Scientist’ role. I messaged him to let him know the outcome and thanked him. However, HG said, ‘that must have been a mistake’ and said he’ll ask the recruiter about it. A few days later, on Aug 30, the recruiter reached out to me to say, ‘Sorry about the mistake earlier! You’ll be in our new grad recruiting pipeline as soon as that position opens up.’ Amonth later, on Sept 28, I received an invitation to the data challenge for their new grad position.

XN, a Senior Applied Scientist at Zillow, is someone Anthony got to know through Twitter during a heated discussion on whether real analysis is really useful for econ grad school. Other than posting on Twitter about my job hunting, Anthony also pinged XN, who gladly offered to chat.

Before talking to XN on Aug 28, I thought the only position relevant for me at Zillow is their econ research team. During the chat, however, I learnt that a possibly even more relevant position is the Applied Scientist role at Zillow Offers, Zillow’s iBuying business. He also said that, even though right now all the positions that are open on that team are for people with experience (i.e. senior roles), he can keep me posted if something relevant comes up. Half a month later on Sept 14, he emailed me saying, a new grad position will open up in 3 days, and asked me if I’m interested, which I said yes to. On Sept 24, he emailed me saying the hiring manager is interested in my profile, and ask if I want to proceed, which I said yes to again. On Sept 25, I had my phone screen, on Sept 30 & Oct 1, the onsite, and a few hours after, the offer.

Something similar happened with LinkedIn — Anthony’s good friend GO introduced me to his good friend JN at LinkedIn on Aug 25. During our chat on Aug 31, he told me that although they have a hiring freeze at the moment, he’d be happy to keep me posted on when things open up again. On Nov 16, JN sent me an email to say they’ve started recruiting again, and sent me a link to a new grad data scientist position. It’s a pity because at that time I’m already at the decision stage with my offers, but if not I would have applied to that position.

These stories are meant to say that, if you have a strong connection with someone, even if they don’t have openings that are relevant for you at the moment, or even if they don’t directly know of anyone who’s hiring, they can ask around for you and/or keep you posted on when future opportunities come up and/or circulate your CV in their org to ask if someone is interested. This is tremendously more useful than ‘go take a look at our career site’.

If you thought referrals are only useful for getting you the interview, you’re wrong — when it comes to speeding up the interview process and/or getting a team match (for firms that do team after you pass the screening, aka Google), referrals again make a huge difference. — more on this in the next section.



Scarlet Chen Senior Data Scientist@Shopify. PhD in Economics@Stanford. Ex-google.