“I could always do data science if academia doesn’t work out.”
The data science hype is real.
It’s a recurring thought many graduate students and postdocs experience, especially if their work involves hearty servings of programming and statistics, the core elements of data science.
Data science can be a rewarding alternative to academia, and academics do have many qualities that make them attractive candidates for data science roles. However, there are also often large holes in academics’ skill sets that can deter them from being hired straight off the bat.
This post will outline the skills needed to make the leap from the ivory tower to industry. We’ll go light on the technical details or business acumen; for a deep dive on those skills, refer to the five-part guide on entering data science.
Especially if you’re just starting to consider data science as a career, it’s highly recommended to think about where your ideal role falls on the analytics-engineering spectrum, which will help identify which skills to prioritize learning.
Where academics excel
Successful research is incredibly challenging. Making it through a Ph.D. (and beyond) requires significant mental and emotional growth. If you’re surviving the challenges of independent research, then you undoubtedly have:
- The ability to distill interesting questions from large amounts of information
- Analytical skills to answer those questions
- Strong attention to detail
- Effective planning and organization
- Resilience to push through (and learn from) failure
Many of the skills gained in graduate school align nicely with what makes for a strong data scientist. In both cases, critical thinking is essential in formulating questions, obtaining data, and deriving meaningful insights from it.
Both professions require effective communication and collaboration with stakeholders from various backgrounds. Additionally, continuous learning and innovative problem-solving are expected in both fields.
It’s difficult to overstate how valuable these skills are. Many senior- or director-level data science positions may require a Ph.D. because of the rigor academic research brings to identifying solutions to complex problems. Below are a few senior roles found with a quick online search, along with their expected education level.
Perhaps these requirements will evolve as the field matures, but currently, having a Ph.D. is a significant asset for career progression as a data scientist.
Where academics struggle
The attractive jobs above, however, assume that much of the mentality from grad school must be unlearned. While the skills acquired in academia are incredibly valuable for data science, their priorities can often be a detriment. Here’s where academics often face challenges when transitioning to roles outside academia.
The speed-accuracy tradeoff
Academia prioritizes precision to the tenth decimal place, as the goal of research is to uncover the truth regardless of time constraints. However, achieving 100% accuracy isn’t always realistic for many companies due to limited resources. Embracing the concept of “good enough” can be challenging, and your workflow may need to shift towards maximizing accuracy within limited time frames.
Implementing the results of an analysis
Unless you’re in a data science role that’s effectively still academia (e.g., Mathematica, Brookings), it’s insufficient to merely generate knowledge. You must also persuade stakeholders to adopt your results, which necessitates an entirely separate set of business skills. If you work for a tech company, integrating your results will also require additional software engineering skills.
These skills gaps can be unexpected and can significantly impact your work. The difference between a mediocre and an outstanding predictive model may require a trained eye, but anyone can see when your model isn’t available to users weeks after you promised it would be.
Focusing on the team over self
Leaving academia means transitioning from maximizing personal output to maximizing the team’s output. While academics invest significant effort in the success of collaborators and students, their name remains prominent on any papers, posters, or seminars produced by coauthors.
Outside academia, you become more anonymous, which can be disheartening. Your contributions may be invisible to anyone outside your company, and even within it.
Likewise, it can be challenging to shift from the independence of academic work to collaborating with coworkers on the same code base, adhering to programming best practices and product management, rather than relying on personal preferences from grad school.
Working on projects you may not prefer
The freedom in academia, especially in U.S. programs, allows pursuing the questions you find most interesting. If you secure funding for your idea, nothing stops you from charting your own intellectual path. In industry, your research direction is often dictated by your organization’s priorities. While you have some say in this, your boss’s decisions can play a substantial role in determining your work.
These challenges take time to overcome, whether adapting to a “100% accuracy” mentality or acquiring software engineering skills like Git, SQL, and API usage. During your job search, having a Ph.D. can be both an impressive credential and a potential hurdle. You may find yourself competing with recent computer science graduates or boot camp participants. The following section will focus on how to navigate this transition smoothly.
Tips for the transition
The mentality
When entering the job market, consider yourself as someone selling your labor to an employer. What skills do you have to offer?
During your Ph.D., you may have developed expert-level skills in specific types of analyses on certain data. For example, you might excel at identifying patterns in astronomical radio waves. However, it’s essential to determine if there are employers willing to pay for these skills. Additionally, consider if you want to continue the same work you’ve done for the past few years.
If you have a deep passion for your research and there are non-academic employers seeking your expertise, you’re already well-qualified for a career outside academia. Congratulations! For others, there is a need to catch up and become competitive applicants.
Regardless of where you aim on the analytics-engineering spectrum, a shift in your self-perception as a coder is necessary for success in the industry:
What you might be thinking: “I can code anything I want.”
What industry wants: “I can code anything anyone asks me.”
In academia, you have the freedom to choose your research questions and methods, which can lead to gravitating toward familiar questions and methods. A Ph.D. often involves becoming highly proficient in a narrow set of skills. However, data science requires a broader skill set, akin to “full-stack” analysts in software engineering who can proficiently code in both front-end and back-end environments, using entirely different languages and perspectives.
Data science is akin to being a full-stack analyst; you must be comfortable switching between extracting insights from data and constructing the infrastructure to convey those insights, such as dashboards and automated scripts.
During your Ph.D., you may have found it easy to stick to the areas of R and statistics that you already knew well. Stepping out of your comfort zone might have been intimidating, as it could reveal gaps in your expertise. You might have felt that others expected you to be an expert in these areas due to your work in your thesis.
Taking formal statistics or R courses might have felt like a challenge to these expectations and exposed areas where you weren’t an expert. Ironically, this fear might have prevented you from developing a well-rounded understanding of R and statistics. In reality, it’s unlikely that anyone cared about your level of expertise in those areas.
Avoid making the same mistake – embrace the idea of not knowing everything and work on filling knowledge gaps. To be an effective data scientist, you need a broad range of skills that should be continuously developed and refined. Consider exploring datasets and challenges on platforms like Kaggle, HackerRank, and Reddit.
Engage in practical projects rather than solely relying on online classes. Hands-on experience will provide a more robust understanding of the topic and give you tangible results to showcase to potential employers.
The transition targets
Unless you are entirely indifferent to your post-academic career path, you may be simultaneously navigating two transitions: transitioning to data science and transitioning to a new field.
The recommended approach is to first transition to data science, acquiring skills in a professional environment where you can learn from your colleagues. Afterward, consider making the transition to your preferred field.
Landing your first job can be challenging, even with a Ph.D., so it’s advisable to cast a wide net. Ideally, seek employment at a company with established teams of data scientists, analysts, and engineers, where you can absorb knowledge from colleagues. This is particularly valuable if your desired field doesn’t already have a substantial number of data scientists and engineers to learn from upon entry.
Once you’ve gained experience in your initial role and feel prepared for your second transition, carefully consider your preferences. In your case, you wanted to contribute to addressing climate change, despite your background in biology.
When searching for jobs, you realized the need for specificity. Did you want to work for a nonprofit, think tank, or government? If government, was it at the municipal, state, or federal level? Did you aim to join a sustainability company, a sustainability department within a larger organization, or a sustainability consulting firm? Furthermore, within sustainability, did you prefer electric vehicles, renewable energy, batteries, aviation, heavy industry, building retrofits, waste reduction, or another focus?
Exact answers to these questions may not be necessary, but thoughtful consideration will help you narrow down your job search. In your case, you decided to focus on technology for sustainability and eliminated think tank, nonprofit, and government positions from your search.
The applications
It’s advisable to remove your publications from your resume. Although it can be a painful step in the transition, the central currency in academia holds minimal value outside the academic environment unless you’re applying to join a team of Ph.D.’s at a think tank. Instead, consider adding a link to your Google Scholar profile, save the changes, and take a break to relax.
While eliminating publications can be challenging, you can put more effort into building a projects section that employers will find more engaging. Writing code is something you can do outside of your regular job, and although it’s controversial and problematic to expect coding side projects on a resume, it’s worth the extra effort when breaking into the field.
Creating a GitHub repository with examples of simple projects, if time allows, can significantly enhance your resume. It provides concrete examples of your coding abilities and problem-solving skills. Think of a hiring manager as someone evaluating an artist – it’s beneficial to see what the artist’s work looks like.
When submitting applications, always include a tailored cover letter that explains why you want to work at the company, why you’d be a good fit, and your significant contributions in previous roles. Extensive guidance on cover letters and resume tips is readily available online. There’s no need for a fancy resume builder; Google Docs offers free one-page templates that work well.
In conclusion
Leaving academia can evoke conflicting emotions. The idealism of academia, where passionate and sharp individuals collaborate to tackle unanswered questions, can be appealing. Being paid to travel to conferences and exciting field sites is undoubtedly enticing. However, a successful academic career demands significant dedication, akin to being an athlete.
Dedication to your research topic, the ability to persist in tackling challenging problems for years, constant justification to funders, and extensive literature review are essential in academia. Additionally, consistent publishing may require relocating every few years, often to unpredictable locations.
Despite academia consisting of individuals who dedicate their days to deep thinking, there is considerable institutional inertia concerning matters such as manuscript vetting, evaluating scientific contributions for tenure, and addressing systemic inequities.
While academia can be an ideal career for some, you may find that data science aligns better with your goals, as you did. Nevertheless, the transition can be challenging, and changing your identity can leave you feeling vulnerable. It’s not uncommon to feel embarrassed or frustrated when taking steps backward in your new career.
You may encounter rejection during your job search, even among peers. I faced rejections from both analytics and engineering teams, despite my background in biology. It’s essential to recognize that the transition can be demanding, but perseverance and a willingness to learn will help you navigate the challenges successfully.
