How I Landed my Senior Data Scientist Job
Part 2: Reflecting and giving advice
This is Part 2 of a series of posts. If you want to see the data behind which jobs I applied to and how I progressed through interviews, check out Part 1 here.
There seems to be a lot of content on the internet about how to land your first data science job, but not about how to make vertical moves from a junior or mid-level position. Six months ago I started a new job as a Senior Data Scientist and I thought others might find it useful to read the story of how I got the job.
This article is not meant to be a “here’s 5 things to do that will guarantee you find a data science job” — far from that. I am telling my story just because I think it will be interesting for people to read about. My nature is always to instruct so I may give some “advice” here and there but most of it is based on a very small sample size (N=1). I’ll list which strategies I purposefully used to find a job versus those that were blissfully coincidental. At the least I hope this article is an entertaining read for a data professional, and perhaps can serve as a source of inspiration for some.
The data
I go into more detail in this blog post, but here’s a graphic summarizing which companies I applied to and how I progressed through the interview processes:
First, some background
(If you read Part 1 you can skip this section)
Hi, I’m Aaron! The end of the story is that I’m a Senior Data Scientist at Saturn Cloud. My role is to help clients utilize our platform as well as create engaging data science content that builds our credibility in the market and educates users about the technologies that we support. This is very different than what a “typical” data scientist (is there even such a person?) does at a company — my role can be thought of as a mix between a solutions architect and developer advocate but geared towards data science. Even so, I believe my experiences here should resonate with anyone looking for a data science job (or any job, for that matter).
I received a PhD in Computer Science from Florida Atlantic University where my research was focused on data mining and machine learning for healthcare applications. Prior to Saturn, I was a Data Scientist at Modernizing Medicine, an electronic health record company. I started my career at Modmed as an intern through a referral from my PhD advisor. This was before they had a data team, and I had the pleasure of growing my career along with the team over several years. Because of this, I was searching for a Senior Data Scientist position without going through a single interview in over six years! I did have a lot of experience interviewing people for data roles at Modmed, but was definitely terrified of being on the other side of the table.
For more background, this was the resume I used for my job search and my portfolio website.
The few things I purposefully did right
There were a couple of things that I consciously set out to do both before and during my job search that I believe helped me navigate through the process and eventually land a great gig!
Keep a short list of interesting companies
I would always be genuinely curious about different data science roles and organizations, and keep a “Companies” note on my phone where I would add any that I found interesting. This helped give me a starting point for when I was actually looking for a job, rather than going straight to blasting job posts on LinkedIn (I did eventually do this, however). It turned that I found the job at Saturn Cloud through LinkedIn, but it was a great experience choosing to apply for companies that I knew.
Dream job checklist
This part is critical to ensuring that a job search is at least somewhat directed. My approach was to send out applications to companies that I was at least mildly interested in, then through the interview process see how much of my “checklist” the role and company covered. This is all about your desires and values: what technologies do you like to use? What industries would you want to work in? What do you value most about a company’s culture? One role probably won’t cover your whole list, but it’s important to evaluate it critically. Otherwise you may take the first offer that comes your way, only to become unhappy then end up doing this whole process over again in six months. This was my checklist, directly copied from the Notes app on my iPhone:
- Established data driven culture, surrounded by smart data people
- High visibility / high impact work
- Doing good
- Work life balance
- Design oriented
Questions to ask employers
This goes along with the previous point, but it’s important that an interview is you interviewing a company as much as it is a company interviewing you. Many of these questions I got from various tweets, LinkedIn posts, or articles from the interwebs. Some of these may not be phrased to directly pose the question to your interviewer, but can also be points that you gather through the experience and research of the company and the team you would work with. These are once again directly copied from my Notes app 😬.
- Where do you see the organization going? Technology? Growth?
- How will I be measured? How can I assure my success?
- How is manager when it comes to feedback? Do you give regular feedback?
- What level of management experience does the manager have?
- Remote work?
- Community involvement — speaking engagements
- Time off?
- Compensation — salary, relocation, time off
- How many employees, revenue, growth plans, funding
- How is your project manager / scrum master?
- Project management methodology / software
- Mac versus PC (PC meaning no-go 😇)
Things I accidentally did right
Making a job search successful often happens way before one starts looking. Here are some things I inadvertently did right before I was even thinking about getting another job. Most of these observations are anecdotal, but I feel contributed to the success in finding my current role. Your mileage may vary!
My network
Okay, this sounds like some boilerplate job searching advice: “build your network!”, “its all about who you know!”. But really, it is. And this wasn’t really something I actively built up because I knew it would help my career, it just came organically from current and former co-workers, attending meetups and conferences, and doing things online.
I was super fortunate to start my career at a high-growth startup and work with some brilliant and amazing people who became close friends. Some of these people went on to work at companies like Twitter, Airbnb, and Facebook. It’s not all about “big tech” companies but it is nice to name drop sometimes 😬. In fact, I applied to several of them and was rejected by all of them (Twitter, Spotify, Facebook, Apple, Tesla — but who’s counting?). Anyway, back to the point. Though these friends are not data scientists, I can still relate to their experiences in the software industry and bounce ideas of them. I was even able to get referrals for roles at Twitter and Spotify (shoutout to my guy Regy Augustin).
The aspect of my network that was most instrumental for finding my current job was the meetup scene. Fortunately, a few data science and machine learning meetup groups popped up in the Miami/Ft. Lauderdale area over the last few years. Miami is nothing like bigger tech hubs such as San Francisco, New York, Austin, etc., but it was great to participate in these communities. It can be pretty tiring going to events after a full day of work (especially when you have to drive from Boca to Miami during rush hour), but I always pushed myself to go because I wanted to meet people and learn things. Many (most) of the times I felt like it was a complete waste of time — the talk wasn’t good, conversations were awkward, wasted gas, etc. But I’m a big fan of this idea:
I’ve found that luck is quite predictable. If you want more luck, take more chances — Brian Tracy
(disclaimer: I have no idea who Brian Tracy is and do not endorse him or his work. I just like the quote.)
You don’t get a “life-changing” encounter at every meetup. Sometimes you don’t even get a marginally meaningful connection. But sometimes you do. And to me, that’s worth trying all the other times.
60% of the time, it works every time — from “Anchorman” movie
(I also do not endorse the product that was being pitched during this scene of the movie)
I probably had a way easier time getting into speaking at meetups because the communities were still burgeoning. I gave my first meetup talk at the Ft. Lauderdale Machine Learning group in 2015 when I was a second-year PhD student. I remember being terrified because the group was super knowledgeable about deep learning, which I knew nothing about at the time. A big break for me was when the Miami Machine Learning group started, and I gave my first talk there in 2018. At the end of a meetup, one of the organizers just said “Does anyone have something they can present at a future meeting?”, and I volunteered. That was it! Eventually that lead to me speaking at several meetups and then taking on an organizer role. Big shoutout to Alex Rubinsteyn who started the group even as a part-time Miamian. Working in New York City, he saw the value that meetup groups provided to the data science community and wanted to bring that to Miami too.
Alex also brought the idea of becoming a PyData chapter and hosting the first PyData Miami conference in 2019. With the work of Alex, myself, and many other organizers, we were able to bring a great event to Miami and plan to have many more in the future. I was even given an intro to a startup through a couple degrees of connections starting from a random Slack message with another organizer from PyData (thanks Ben Suutari!). Saturn Cloud, the company I now work for, was actually sponsor of the PyData Miami 2020 conference (that was eventually cancelled because of you-know-what). I didn’t even put two and two together until I was halfway through the interview process — it turns out I already had several emails with my interviewers months before I applied 😅.
Public-facing work
This is a big one, and something many people give as advice for data scientists looking for a job. Saying you know a tool or algorithm on your resume doesn’t give the recruiter any context or depth around your knowledge of the technology. GitHub repos, blog posts, and talk recordings give recruiters a chance to see what you really know. This can be challenging sometimes when doing data science work at a company, because you are not able to make your code or data public for others to see.
Fortunately for me, I was able to give a talk at the Spark+AI Summit in 2018 related to my work at Modmed (of course after going through several reviews by the legal and marketing teams). I also gave numerous talks at conferences, meetups, and miscellaneous events that were related to my PhD research and other side projects. I always made sure to record the talk and post it on YouTube, even if it was just a screencast from my laptop.
Though I hadn’t contributed to an open source project before this job search, I did have a public GitHub repo with code that I wrote to conduct machine learning experiments for my PhD research. I actually ended up doing a presentation about this framework during my interview process for Saturn.
Doing lots of different things
I had a few very different companies tell me that I would be a perfect fit for a role, and I think this is because of the variety of work I’ve done and published throughout my career. I found myself covering work across both the data engineering and data science worlds, and my PhD research allowed me to dive deep into machine learning theory. At the very least, being interested in different topics and reading up on them gives you a starting point for conversations. Sometimes you just need one hook to get an “in” with your interviewer!
Well designed portfolio
I don’t have any quantitative measure of how effective my portfolio site was, or if recruiters even looked at it. But I believe that a well-designed and eye-catching portfolio is something that sets data scientists apart from the pack or at least makes your application more memorable for a recruiter. I don’t take any credit for the UI design (thank you Camila O.H.), I just put all the <div> tags in the right place.
If anything, the site served as a way for me to brush up on my HTML/CSS/Javascript skills (mostly reminding me why I chose data science, no offense to web developers 😅). But it also helped me review the work I had done so far in my career, keeping everything fresh in my mind when it came time for interviews.
COVID-19
This is certainly nothing I did, but I do have some observations about how COVID-19 lockdowns and remote work affected by job hunt. This was (and is) a terribly difficult time for many people, and I am extremely fortunate to be in an industry that was still hiring during those uncertain times. Working remotely came to my advantage when scheduling interviews as I didn’t have to figure out a way to sneak out of the office to answer a call, or take a random day off to travel to an onsite interview. Being stuck in my aparment on evenings and weekends also helped me reflect more on what I wanted in the next job and spend time curating my resume and portfolio.
What I learned
The job search was definitely an enlightening experience for me. I did not anticipate the mental and emotional toll it takes on you, from pulling late nights filling out applications, to being crushed when your “dream job” rejects you, to having to be in top-shape for a demanding technical screening — all while still maintaining the responsibilities of your current job and everyday life!
Whiteboard coding interviews are difficult
This could be a whole blog post in and of itself. I only went through a couple FAANG-style coding interviews (actually it was just F) and they were miserable. I went through the software engineer and data science interview process for Facebook, and while they were a bit different, I froze up during both of them. Looking back, I know that I am fully capable of answering those questions, but the environment wasn’t really conducive to thinking clearly.
For counter-examples, I had take-home coding tests from several smaller companies. They were all well-defined problems, and I had time to think about them and work on them in the same manner I would on the job. They also served as examples of what my work on the job might be like — I ended up withdrawing a couple of applications before completing the test partly because I was not interested in the work!
Having an offer expedites you through the pipeline
This was no means planned, but when I was expecting an offer from Saturn Cloud I was in the middle of the interview pipeline with a few companies. I wanted to make sure that I got the chance to evaluate those positions so I told the companies that I had an outstanding offer, and that really sped things up. With one company I had an initial phone screen with a recruiter, mentioned that I was interested in the job but had an offer coming, and they scheduled an interview with the Director of Data Science on the same day!
Interview pipeline speed gets good candidates
This is reflection more for a company hiring than for someone looking for a job. Many of the rejections I received came after I had already begun or completed interview processes with other companies. I specifically recall one of my interviewers at Saturn Cloud telling me that their goal was to move quickly through the interview process, and for it to not take more than 6 hours of my time total. I really appreciated this and it spoke well of the company — that they thought through the interview process and respected my time even before I was an employee there.
That’s a wrap!
For anyone that made it this far — thanks for reading! I hope this story was interesting and helpful for anyone looking to level-up their data science job.
Feel free to contact me on Twitter @rikturr to keep the conversation going!
