From my experience, people who work in Data Science have a wide diversity of backgrounds. I (like many of my fellow data scientists) didn’t start my career in data science right out of college. I started out working as a securities broker in the investment finance industry. I quickly discovered that the career path I originally chose was not a good fit and started a multi-year journey towards becoming a data scientist. In this article, I’m going to share 8 strategies I used in my successful transition to becoming a data scientist. Let’s get into it!
Strategy 1— Know what you want
Data science is a competitive industry and it can be difficult to get into, especially if you weren’t originally planning on it. It is crucial that you know that you really want to work in the field – if your journey will be anything like mine, it will take some significant time and effort to break into your first data science job. You need to be sure this is what you want so you stay focused and don’t lose motivation in your journey!
Strategy 2 — Set up a job feed (even if you are no where near qualified to apply)
When people ask me for advice on becoming a data scientist, I always say this one first! Set up a daily job feed for data science positions and record key details from the job postings. This job feed isn’t for applying, it is for learning what employers want out of a data scientist. I could give you a list of what I think you should learn, but I’m one person with one opinion. If you have data from 50+ job postings, you don’t have a bunch of opinions, you have real information about what real employers want from real data scientists!
When I first decided that data science was the goal, I set up a daily job feed for any data science positions in the Dallas area – making it location-specific was important because it allowed me to not only learn what employers want, but it helped me create a list of target companies to focus on. Every day I got an email, usually with 2–3 new job postings. I wasn’t anywhere near qualified for any of them, but again applying wasn’t the goal, data was the goal! I saved the details for every job posting in an excel spreadsheet – this gave me extremely valuable data to direct my journey.
The job feed gave me these key pieces of knowledge:
- Skills that employers want from data scientists – programming languages, specific ML algorithms, statistical knowledge etc.
- Specific tools that data scientists use – Python, SQL etc.
- Educational background — what level of education in which subjects
- Companies that hire data scientists in my area (and an idea of how large their data science groups are based on the number of postings I saw)
- Occasionally I would see salary ranges – this wasn’t very actionable information, but it was good for keeping up my motivation since the posted salaries were a lot higher than mine 😅
Without the job feed, I would’ve had to rely on blogs, articles and pieces of advice from other people. The job feed gave me a list of what employers in my area wanted and the source was the employers themselves!
From this process, I made my ‘to-learn’ list based on the skillsets that I saw most frequently on the job postings. That list became a roadmap for the rest of my journey to becoming a data scientist.
Strategy 3— Try to gain some data science skills in your current job
Different jobs have different levels of flexibility in the work you do and how you do it. If you have some flexibility, try to do a few things that will help you gain skills from your skillset list (from strategy 2).
Example #1 from my journey:
I realized I wanted to work in data science while I was working at Fidelity doing mutual fund operation work. My work was pretty well-structured (meaning not a lot of flexibility) but, I was able to carve out some time for ‘pet projects.’ For one of my projects, I built a simple linear regression model that predicted the number of data discrepancies we would see based on the daily market volatility. It really wasn’t much, but it provided me a tiny amount of "data science" experience that gave me (1) more confidence that I wanted to work in the industry (because I loved making that little model) and (2) one line that I could put on a resume that demonstrated (in a very small way) a data science skill.
Example #2 from my journey:
I later transitioned from Fidelity to GM Financial (I’ll talk about the strategy of that move in the next section). After working at GMF for a while, I decided it was time to start gaining Python experience (which I had as a pretty high priority on my skill list from the Strategy 2 section). I asked my manager if I could download a license free version on Python and to my surprise, he said ‘no’. I protested and explained that it was free and that I could use it to help with my job responsibilities, but the answer was still a firm ‘no’. I decided to start looking for other jobs because of that. I know it seems like a small thing, but remember, my goal was to become a data scientist, not to keep my current position. I needed some Python skills to accomplish my goal! I got a job offer from a company where I would be able to use Python. When I told my manager about it, he asked me what he could do to get me to stay — I just said I wanted to work on a project or two in Python since it is in alignment with my career goals — this time he said ‘yes’ 🤷♂️! I then worked on some small modelling projects in that position, which gave me valuables skills and resume talking points.
Strategy 4— Be willing to do intermediate jobs to gather the skills needed
Ideally, you can just jump from your current, non-data science job, directly into a data science role. For my journey however, the skill gap between what I had and what I needed was too large for just one jump. Because of the size of the gap, I had to take a couple of different roles to stair step my skills. It can be hard to switch jobs for skills, but if you really want to be a data scientist, you may have to pursue this strategy.
I had two ‘intermediary’ roles in my journey towards data science. Each role I used to gather a subset of skills I needed to become a data scientist or I at least needed for the next role.
- Pricing Analyst (Fidelity) – This job got me off of the phones as a broker and into a more analytical role. The main skill I had was working with data in Excel – that skill isn’t super data science friendly, but it gave me enough to get into my next role, where I picked up more skills!
- Data Analyst (GMF) – I was able to qualify for this job because of my financial background and my excel skills. In this role, I picked up heavy data analyst experience, SQL experience, and experience using an old-school statistical program called SAS. And of course, as I mentioned in Strategy 3, I picked up just a little bit of Python skills even though it wasn’t a part of the job description.
Strategy 5— You might have to go back to school, do it part-time!
During my data science journey, I started a master’s in data science at the University of Oklahoma. I think that having this on my resume really helped me get the interview that ultimately gave me my break into data science. I found that most data science jobs required at least a master’s degree (something I learned from Strategy 2) – so I decided that I would work on getting that requirement, but I would do it part-time.
Pursuing education part-time was one of the best decisions I made during my journey. It balances work experience and education (and you get to have a salary while studying, which was really nice compared to my undergraduate experience 💰 ). It took me an extra year to get my degree, but I continued to gain valuable experience and I could start working as a data scientist before I finished. I think now, with the taboo of online learning almost completely gone (or at least much lower than it has been), there is no reason you need to quit your job for a master’s degree. Get your cake and eat it too by getting work experience and education at the same time (oh and avoid those student loans as well)!
Strategy 6— Teach yourself, don’t rely on schooling
I learned a lot from my master’s program, but in reality, it taught me a pretty small subset of what I needed to know to become a successful data scientist. I also couldn’t wait around for my program to teach me the skills I needed, I didn’t have the time!
I had my first data science interview when I was just a few classes into my master’s program. Thankfully I had taught myself a pretty wide array of data science knowledge in the years leading up to the interview. The interview was fairly intense with a long case study that required a general understanding of multiple machine learning and data science topics. I relied 100% on my self-taught knowledge in the interview – which went well enough for me to secure the offer.
With the huge amount of resources available online today, there is no reason to not teach yourself. I did learn from my master’s program, but I estimate that about 85% of my data science knowledge comes from self-teaching. Going back to strategy 1 from the beginning of the article – if teaching yourself data science is fun and interesting, you can know that it is a good career path for you.
Strategy 7— You were doing something before data science, use that domain knowledge to your advantage!
If you don’t have data science work experience (like I didn’t), you can use the domain knowledge from your other work experience to edge out some of the competition. This was a really important factor for me getting my first data science job.
While I was working at Fidelity, I took and passed the Chartered Financial Analyst tests (CFA). This is a pretty difficult designation in investment finance. My first data science position was at Toyota working on the financing side. Although it wasn’t the exact same type of ‘finance’ (consumer vs. investment), the certification showed that I had a level of professional financial knowledge.
The CFA helped, but what gave me the biggest edge was the fact that I had industry experience in a pretty niche area, i.e., "captive auto finance." This is a very specific industry made up of consumer finance companies that are wholly owned by a manufactorer and whose purpose is to originate loans for the purchase of the manufactorers’ products. When I landed my first data science job, I was working at GM Financial (captive finance company of General Motors) — My first data science job was at Toyota Motor Credit Company, which is the captive finance of Toyota. In my interview I was able to ‘talk shop’ with my future boss, using terminology that only industry insiders would use. I don’t know for sure, but I bet that given the industry work experience, I could have beat someone that had a little bit of data science experience, but no industry experience. I do know that it really helped!
The takeaway here is, focus on applying to data science jobs in an industry you’ve already worked it. This could help compensate for your lack of data science experience and can give you a competitive edge over applicants with some data science experience that are coming from other industries.
Strategy 8- Be willing to take risks; a successful journey may require it
To become a data scientist, you may have to take a few uncomfortable risks. If you really want it (again – and you are probably annoyed at me by now for this – going back to Strategy 1), taking reasonable risks are worth it!
The big risk I took
I took multiple risks in my data science journey, but the biggest one was when I finally got my first ‘data science’ offer.
The first opportunity I had to become a data scientist (at Toyota) was a 1-year contract position. This was a significant risk to me, because I was leaving a reasonable paying full-time job for a 1-year role that could get extended or convert into a full-time position or it could not! Contracting was probably the only data science offer I could get at the time. Being willing to take a contracting position gave me two advantages (1) fewer people are willing to take contracting positions – I was competing with a smaller applicant pool and (2) less risk on the side of the hiring manager, making them more willing to extend an offer – if things didn’t work out, it was easy to terminate the contract early, or just let it expire at the end of the year. Since I was taking most of the risk, they were more willing to hire me without data science experience. I had to bet on myself, and it was a risky bet! Thankfully (and a lot of thanks to my manager at the time) I was later able to be converted to a full-time position, the bet paid off!
If you aren’t a data scientist, but want to become one, you may need to take a few risks to keep your journey’s momentum going. I suggest you take reasonable risks as needed, it may be required for you to reach your goal!
Conclusion
Breaking into data science was difficult and intense endeavor that required some clever strategy, time and a lot of work. But for me, the journey was well worth it! I’ve now been working at Toyota for nearly six years as a data scientist and it has been fantastic! I started out as a contractor with no data science experience and was able to progress to managing a team that tackles big and impactful problems! Of course, it would be arrogant to take credit for my successful journey, I had many mentors and great managers help me along the way. I also had some good luck as well – the contribution of luck to anyone’s success can’t be ignored 😊 .
I hope that this was helpful for you and, if you are looking to break into data science, I wish you the best of luck!
