After being shot, I received four job offers in data science and more than doubled my income in just two months


Many people’s careers are being impacted by the pandemic covid-19 at this unprecedented moment. This group includes some of the best data scientists I’ve ever worked with. I felt it was worth sharing publicly after sharing my personal experience with a few close friends to assist them in finding new jobs after being laid off. After all, this affects more than just my friends and me. Any data scientist who has been laid off as a result of the pandemic or is actively searching for a data science job will find something here to relate to, and I hope it will give you hope in your job quest.

So, if you’ve ever been lost – in getting interviews, interview planning, negotiation, or something else – I’ve been there and I’d like to assist you. If you think I might be able to help you in some way, please contact me here! Here’s how it went down for me. I hope you find some helpful hints and inspiration in it.

Getting Dismissed

My boss told me in December 2018 that I will be laid off in January 2019. My then-startup company’s VP of Engineering had written a letter to our head of People Success three months ago. This letter outlined why I was one of the company’s best performers and argued for a raise in my pay. This resulted in a 33 percent pay raise for me. I was naturally pumped and excited to reach the next milestone on a significant project. The future of the business, as well as my own, appeared to be promising. During this triumphant moment, I was informed that I had been affected by the company’s cost-cutting initiative. On January 15th, I was shot.

To say the least, being forced to start searching for a new career was terrifying. After looking through the available data science job openings, I quickly realized my knowledge gap. Many of the work specifications out there, such as product sense, SQL, stats, and more, were simply unrelated to what I was doing at the B2B startup (a combination of entry-level data engineering and machine learning). I knew the fundamentals, but I didn’t know how to bridge the gap to more advanced skills. Even that problem, however, seemed to pale in comparison to more pressing concerns, such as how do I even get an interview? I had just 1.5 years of experience working for a startup and no background in statistics or computer science. Soon after, further concerns emerged. What if I can’t find work before my visa expires? What if the economy tanks before I’m able to find a new job? Despite my doubts, I didn’t have much of a say. I needed to look for a new career.

Getting Ready for the Search

I needed some facts to determine my next move in the face of what felt like an impossible mission. After doing some analysis, I discovered that more than half of the data science positions on the market were product-driven (‘product analytics’), with the remaining positions focusing on modelling or data engineering. Other than product analytics, I noticed that positions with higher requirements appeared to be more common. Most modelling jobs, for example, required a PhD, while engineering jobs required a computer science background. Because the specifications for different tracks were clearly different, it was only natural that the training for each would be different as well.

With this information in hand, I made a critical decision: planning for all tracks would be exhausting and ineffective. I’d have to concentrate on only one. I chose product analytics because it was more likely that I would get interviews on this track based on my history and experience. Of course, not everyone in data science has my specific background and experience, so I’ve outlined the general criteria for three types of data science jobs at large corporations below. This easy rundown saved me a lot of time, and I’m confident it would be helpful to those looking for data science jobs. I would note, however, that for small businesses, the interview can be less formal and involve a combination of all three.

Product Analytics (~70% on the market)

  • 70 percent of the industry is devoted to product analytics.
  • Data Scientist, Analytics at Airbnb; Data Scientist at Lyft; Data Scientist at Facebook; Product Analyst at Google Modeling (20% on the market)

Modeling (~20% on the market)

  • Knowledge of machine learning (not only how to use it, but also the underlying math and theory) is needed, as well as strong coding skills.
  • Data Scientist, Algorithms at Lyft; Data Scientist, Algorithms at Airbnb; Applied Scientist at Amazon; Research Scientist at Facebook Data Engineering (ten percent of the market) are just a few examples.

Data Engineering (~10% on the market)

  • Requirements: a computer scientist with end-to-end data engineering skills; understanding of distributed systems; MapReduce and Spark; practical experience with Spark; good coding ability
  • Examples include a Data Scientist at Airbnb’s Foundation and a Data Scientist at a few startups.

The remainder of this article is geared toward those training for product analytics roles, based on my own experience. Check back later to see my article on how to prepare for a data engineering job.

The Job Seeking Process Begins

When I found out I was going to be laid off, the first thing I did was apply for as many positions as possible. I applied for jobs on all of the work boards I was familiar with, like GlassDoor, Indeed, and LinkedIn. I also sought referrals from everyone I met. However, since it was almost the end of the year, I didn’t hear back until January 2019.

Requesting referrals was much more successful than applying on my own. I only got three interviews out of around 50 raw applications, but I got seven out of 18 referrals. Overall, it was becoming clear that I was not a good candidate for this position.

Overview of the Interview

Although each company’s interview structure differed, there was a general outline that most companies followed:

  • An initial phone call from a recruiter
  • TPS (technical phone screen) (one or two rounds) or a take-home assignment
  • Three or four rounds of technical interviews and a behavioral interview with recruiting managers are usually included in a four-to-five-hour onsite interview.

A take-home assignment was given before or instead of a TPS at about half of the companies I interviewed with (4/10). Take-home duties took a lot of time and effort.

After submitting an 8-hour take-home task, I usually required at least a half-day of rest. As a result, I did my utmost to arrange the interview in a timely manner. The morning after my take-home assignment, there were no interviews scheduled. Simply understanding the fundamental structure will help you feel more at ease and capable of coping with the job search process.

Prior to the Interview:

Every chance was crucial to me going into my interviews. Despite the fact that I was aware that certain people learn by interviewing, improving after a number of interviews, and usually receiving offers for the last few companies for which they interview, I did not believe I was capable of doing so. Out of 500 raw applications, I only got four interviews when I graduated in 2017. I didn’t anticipate receiving several more in 2019. As a result, my strategy was to be fully prepared for and interview. I will not let an opportunity pass me by.

One of the advantages of being laid off was that I could devote all of my time to studying for the interview. I organised my studies per day, concentrating on two or three topics a day. That is no longer the case. I had learned from previous interviews that having a comprehensive understanding helps you to provide more detailed answers during interviews. In an interview situation where you are more nervous and anxious than normal, having a depth of knowledge is particularly beneficial. Now is not the time to try your hand at impersonation.

As I explain my own experience, I can’t help but think of a common misunderstanding I’ve heard: it’s impossible to acquire product/experimentation expertise without firsthand experience. I completely disagree. I had no previous experience with product or A/B testing, but I assumed that by reading, listening, thinking, and summarising, I could learn those skills. After all, we were taught to think in this manner in school. In fact, as I meet more senior data scientists, I’m learning that this approach is popular, even among those with decades of experience.You may not be asked about what you were doing at the time, but you can obtain the insight you need in ways other than work experience.

Here’s a quick rundown of what to expect. During a TPS, product and SQL questions were frequently asked. A few rounds of questions were asked during onsite interviews, including product knowledge, SQL, statistics, modelling, behaviour, and possibly a presentation. The following parts list the most valuable tools (all of which are free) that I used while preparing for interviews. GlassDoor was a reliable source to get a sense of company-specific issues in general. After I recognised the issues, I was able to determine both what the organisation required and where I fell short in meeting those requirements. I was then able to devise a strategy for closing the holes.

Being Ready for Specific Subjects

The six subsections that follow describe how I prepared for the particular material that comes up in product analytics monitoring interviews. I hope that by discussing my own training, I can smooth the way for those who come after me.

Sense of the Product

I was primarily responsible for designing and implementing machine learning models and writing spark jobs while working as a data scientist at a startup. As a result, I learned very little about the product. I had no idea how to approach real interview questions like “how to calculate success?” or “how to validate the new functionality through current users’ behaviours?” when I saw them on GlassDoor. They seemed much too vague and open-ended at the time.

To develop a product meaning Using the tools mentioned below, I resorted to the simple read and summarise strategy. All of this reading helped me gain a better understanding of the product. As a result, I devised a method (my own ‘framework’) for responding to any sort of product query. I then put my knowledge and structure to the test with practise, which is crucial to mastering any ability. Answers to product sense questions were written down. I practised my responses aloud (even capturing them on my phone) and used the recordings to improve them. Soon, I wasn’t just faking it for an interview; I really knew what I was talking about.


Stellar Peers
Cracking the PM Interview by Gayle Laakmann McDowell and Jackie Bavaro
Decode and Conquer by Lewis C. Lin
Case Interview Secrets by Victor Cheng


I failed a SQL TPS the first time I took one, and it was with a business I was really involved in. There was clearly a need for improvement. I wanted to practice once more, so I spent some time drilling SQL queries. I eventually completed questions that had previously taken me a week to complete in a day. It is said that practice makes better.


Leetcode database problems
HackRank SQL problems

Probability and Statistics

I did some coding exercises and reviewed basic statistics and probability to prepare for these types of questions. Although this may be daunting (both topics have a tonne of content), the product data scientist interview questions were never difficult. The tools mentioned below are excellent for reviewing.


Khan Academy has an introductory Statistics and Probability course which covers the very basics of both.
This Online Stat Book covers all the basic statistical inference.
Harvard has a Statistics 110: Probability course which is an introductory course on probability with practical problems. If you prefer reading to listening, PennState has an Introduction to Probability Theory course with many examples.
I also coded through 10 days of statistics on HackRank to solidify my understanding.
Sometimes, A/B testing questions were asked during a stats interview. Udacity has a great course to cover the basics of A/B testing and Exp Platform has a more concise tutorial on the topic.

Machine Learning

I went into the job hunt with little machine experience because I didn’t have a computer science degree. I had taken some courses at my former job, and I went through my notes to prepare for the interviews. Despite the fact that modelling questions are becoming more common, product data scientist interview questions are mostly focused on how to implement certain models rather than the underlying math and theories. Even, here are some tools to help you brush up on your machine learning skills before your interview.


To start I recommend this free Applied Machine Learning course by Andreas Mueller
Coursera – Machine Learning by Andrew Ng
Udacity – Machine Learning Engineering Nanodegree


Some employers wanted applicants to show either the take-home assignment or a project that they were particularly proud of. During behavioral interviews, other organizations inquired about the most impactful project. The key, regardless of the format, is to make your presentation engaging and challenging.

That sounds fantastic, but how do you go about doing it? My key suggestion is to consider all of the specifics, from high-level targets and performance indicators to ETL, modelling implementation details, deployment, tracking, and progress. Rather than one huge idea, the little items add up to make a better presentation. Here are a few questions to consider in order to achieve your ideal presentation:

  • What were the goal and the success metric of the project?
  • How do you decide to launch the project?
  • How do you know whether customers are benefiting from this project? By how much?
  • How do you test it out? How to design your A/B test?
  • What was the biggest challenge?

You want to involve the audience while introducing a project. To keep my presentations engaging, I often share interesting results and the project’s biggest challenges. However, practising is the best way to ensure that you are participating. Train aloud as much as possible. I practised giving presentations in front of my family to ensure that I understood the content and could communicate effectively. An interviewer who is required to listen has no chance if you should engage the people you meet.

Behavioral Question

Although it’s easy to get wrapped up in planning for technical interview questions, don’t overlook the behavioural interview questions. During the onsite sections of every company I’ve interviewed with, there was at least one round of conduct interviews. Usually, these inquiries fall into one of three categories:

  • Why us? / what do you value most in a job?
  • Introduce yourself / Why are you leaving your current job?
  • The biggest success/failure/challenge in your career. Other versions: Tell me about a time you resolved a conflict or you’ve had to convince your manager or a PM on something.

For data scientists, behavioural problems are crucial. So, be ready! Understanding a company’s mission and core principles aids in answering the first set of questions. Telling a storey can be used to answer questions 2 and 3; three stories were sufficient to answer all behavioural questions. When you go in for an interview, make sure you have a few good stories ready. I practised a lot by saying it out loud, recording it, and then listening to fine-tune my responses, much as I did with product questions. The only way to ensure that a storey succeeds is to hear it.

The Secret to Getting a 100% Onsite-to-Offer Ratio

It was usually a busy, hectic night the night before an onsite interview. I tried to cram in as much technical information as I could when reviewing my statistics notes and formulating my framework to respond to a product query. Of course, none of that was particularly useful, as we all learned in school. The results were primarily dictated by the amount of training done prior to the exam, which did not include a single night of cramming. So while planning is important, there are some guidelines you may follow on the day of the interview to ensure a successful outcome.

  1. Before responding to a comment, make sure you’ve clarified it. By repeating the question in your own terms, you can make sure you understand what you’re being asked. If you answer the questions without clarifying them, it’s a red flag.
  2. Organize the responses to all of the questions. Make a list of your thoughts in bullet points. This demonstrates to interviewers that you have a methodical approach to solving a problem and aids interviewers in writing a critique for you later.
  3. When you don’t know the answer, don’t freak out. It’s fine if you’ve never heard of the domain before. In such cases, you might begin by making some assumptions, but make sure to communicate that you’re doing so and ask if they’re fair. It’s perfectly acceptable to request more time on occasion. What if you’re at a loss for words and your mind is blank? Discuss a personal experience that relates to the question.
  4. It’s all about how you feel. Companies are searching for people who are able to listen and accept various points of view. You want to demonstrate that you are a pleasant person to work with. Be respectful and humble. Pay attention and ask questions. Bring your positive energy into the room, and make an effort to engage in meaningful conversation.
  5. Investigate the company. Get to know the company’s goods. Consider how you can develop the goods and what measures you can use to assess their performance. Reading the blogs of each company’s data scientists will also help you understand what they do. This type of research leads to more in-depth and, eventually, better interviews.

The following is the input I received from onsite interviews using these rules:

  • Answering product questions in a very organized manner
  • The presentation is well-organized and well-planned.
  • Showed a keen interest in our goods and provided helpful suggestions for improvements.


The next move was to coordinate with recruiters to finalise the numbers after receiving verbal offers. There’s only one law I follow in this situation: ALWAYS compromise. But how do you do it?

During my offer negotiation process, I followed Haseeb Qureshi’s very helpful guide on negotiating a job offer (complete with scripts!). Every single rule was absolutely right. I bargained with any company that made me a bid. The average rise in deal value was 15%, with the highest offer increasing by 25% in total value. Negotiating is successful, so don’t be afraid to give it a shot!


  1. The trick is a lot of work.
  2. Failure is a part of life, and it’s also a part of job hunting. Don’t get too worked up about it.
  3. Find a stress-relieving method that works for you.


I eventually got four work offers within two months of being laid off, after losing 11 pounds and a lot of cries and crying (job searching is stressful, and it’s okay to admit that). Three of the offers were from companies I had never considered joining: Twitter, Lyft, and Airbnb (where I eventually accepted), as well as a healthcare startup. I had a total of 10 interviews, 4 onsite interviews, and 4 work offers by the end of two frantic months, giving me a 40% TPS-to-onsite rate and a 100% onsite-to-offer rate.datasciencejobfunnel

Timeline from being laid off to joining my dream company (photo by Emma Ding)

After being laid off, I was extremely fortunate to receive a great deal of support and assistance from family and friends, which was crucial in helping me secure a job at my dream business. It was a challenge. Ironically, searching for a career is a lot of work as well, but it was all worth it in the end.

Since I was overwhelmed, I decided to write this blog. There is a lot to think about when it comes to interview preparation. I hope this post has clarified things for other data specialists looking for jobs, and if you have any questions, please contact me here. I’m lucky to be employed in a wonderful career right now, and I’d be able to assist you in getting there as well!