12 Best Practices to Grow in Your Data Career

A career in data can expose you to a wide range of technologies, skills, and stakeholders. You’ll have the opportunity to work on projects with a significant impact across core areas of a business, often within highly cross-functional teams or in collaboration with other technical colleagues. 

The move from junior to senior data science roles will look very different depending on your industry and specialty, but for most data scientists this takes 2-5 years. Salaries can easily reach the 6-figure mark with the right knowledge and skills. In 10 years' time, demand for data scientists is expected to increase significantly, although the role may look very different, as new tools, best practices, and methodologies are still emerging.

Growth in a data career can take many forms, but whatever your goals are, there are some essential tips and practices to keep in mind.  Read on to learn our best practices to ensure you are able to make consistent progress throughout your career.

1. Define Success

Working for work’s sake isn’t a problem in itself; continuous practice is regarded as a good thing and many feel the benefits. However, you may not feel satisfied if you don’t have a goal in mind to grow towards. It is essential to sit down and define what success looks like for you. Not only will this ensure you are moving in the right direction, but it will also give you a way to evaluate your progress over time. Your idea of success doesn’t have to be about promotions, income, or job titles. You might aspire to long-term goals such as publishing a paper or completing a Ph.D. Perhaps you want to lead a team or develop a product used by millions. By defining your vision of success, you can identify the necessary steps to take to grow in the right direction.

2. Don’t Stop Learning 

Have you ever heard anyone talk about their growth mindset? Research shows that those with this mindset tend to achieve more than those who believe talents are innate gifts. Data careers can be tough; many practitioners will face setbacks and encounter failures. Learning from mistakes is often cited as one the key metrics for success; even successful entrepreneurs will talk about their many failures and what they learned from them. 

Another key reason to continuously develop your skills is to keep up with the fast pace at which the industry is changing. Elad Cohen mentions in DataFramed (around 49 minutes in) that regardless of whether or not you think automation is going to streamline away parts of the data science workflow, many people tend to underestimate what technology will look like in 10 years. If you don’t keep learning and growing, the industry itself may move faster than you can.

3. Keep Building A Portfolio

The importance of a portfolio cannot be understated. Although they are often associated with junior data scientists trying to land their first role, a portfolio can be a great place to collect some of your best work over your career.  A portfolio can also be an outlet to work on topics that you are passionate about, particularly if you don’t get the opportunity to do this day-to-day. Finally, expanding your portfolio can motivate you to apply the previous best practice of always developing new skills. 

Maintaining a portfolio throughout your career ensures you can:

• Share your work if you are moving roles, companies, or even going freelance
• Capture some of your most interesting work to build on later
• Have a space to share outputs of your passion projects, which other data scientists love to see
• Build up your own personal brand, which can open doors in other areas of your career

4. Invest In Model Deployment Capabilities 

At the moment, one of the most significant disconnects between how data science is taught in university and how it’s applied in industry is MLOps. In a nutshell, MLOPs is a set of tools, practices, techniques, culture, and mindset that ensure reliable and scalable deployment of machine learning systems. One of the biggest challenges facing data teams today is the scalable and efficient deployment of machine learning models, which is why estimates show that as low as 8% of enterprises have models in production. It is no surprise then as to why MLOps is one of the fastest-growing areas of data science in the past year.

Investing in model deployment capabilities is essential for many data scientists today, even those with years of experience. Recognizing that there may be a gap in your knowledge and identifying steps to remedy it is a great way to grow. You don’t have to have expert knowledge in every new tool,  but understanding the core concepts of MLOps or other novel developments and learning how to apply their techniques will get you a long way and lay the foundation for specialization later. If you’re interested in getting started with MLOps, you can find some free resources to get started in this blog post.

5. Choose A Specialty

Data science is an exceptionally wide field. In this e-book, DataCamp defines data science as “a cross-disciplinary field that seeks to extract meaningful insights from data”.  The different disciplines can range from mathematics, to statistics, machine learning, programming, data visualization, and more. There is no way one person can develop expert knowledge in all of these areas, particularly when significant knowledge is needed to implement techniques in a single area. For instance, consider how much theory you need to understand in order to deploy a natural language model while achieving satisfactory accuracy and avoiding bias.  

Once you have a little experience in a wide range of areas, it’s worth your time to consider where you might want to specialize. Becoming an expert in one area is a great way to open doors and bring you closer to your vision for success.

6. Build Out Your Network  

Networking is one of the most important ways to grow in your career, and it can come in many forms such as attending events, setting up “coffee chats” with people in the industry, or coworkers, or even giving talks yourself. A network is a great way to find out what is going on at other organizations and what new skills you might need to learn or may be needed at your organization in the future. It’s also the single easiest way to get your foot in the door at another company if you’re looking to change jobs.

Many people find the idea of networking to be daunting, but there is a range of ways to network and find events in the data community. We collated here a list of top data science conferences to attend in 2022 to get you started on your networking journey. 

7. Combat Imposter Syndrome

Imposter syndrome is defined by Harvard Business Review as “doubting your abilities and feeling like a fraud,” and it's something that many data scientists have felt at some point in their careers.  The breadth and depth of knowledge that experienced data scientists have can seem like an exceptionally high bar to reach in the early stages of your career. Even later on, you’re likely to encounter experts with exceptionally deep knowledge in each of the many subfields of data science. Imposter syndrome is a term that has become much more common over the last few years, with many high-powered executives opening up about their experiences. As such, there are many books available to help understand it might affect you and what you can do to combat it. This Reddit post also contains several data scientists sharing their experiences with imposter syndrome.

8. Keep An Eye On The Job Market

Career growth is highly correlated with how you’re spending your nine to five. If what you do every day isn’t aligning with your long-term definition of success, you may want to re-evaluate your position. This could mean having a conversation with your manager to change what your day-to-day looks like or making a lateral move within your organization. However, it is always worth understanding what is happening in the job market. You may find that your skills are in high demand, and that other companies can offer different roles or wage and benefit increases.

Jerry Lee from Wonsulting, a career development firm,  recommends interviewing every year at companies you are interested in order to keep your interviewing skills fresh and ensure your current company values you. The importance of this depends on your long-term goals and success criteria, but keeping your finger on the pulse of the rest of the market is always a great idea.

9. Stay Up To Date

Aside from developing your own skills, it’s also important to have an overview of what is happening on the cutting edge of research, tech companies, and market leaders in your space. You don’t always delve into the details—understanding the potential of the latest tech at a high level is a great way to stay informed. Bonus points if you’re able to have conversations with your network about this tech to either learn how they’re applying it or teach them something new!

A good resource for this type of high-level scanning is Adam Votava, who writes a weekly “Keeping Up with Data” post on LinkedIn. These well-written summaries are a great way to keep tabs on the latest research, business news, and even recently released Python packages. You can also subscribe to our Weekly Roundup video newsletter, which provides you with the most interesting news coming out of data science, tech, and the research communities. 

10. Consider The Impact Of Your Work

Growth in your career should lead to meeting your own metrics for success – but what about your organization’s success? It is important to consider what the impact of your work has been, whether it’s saving time through efficient tools, helping coworkers by creating an informative dashboard, or saving on production costs with a well-trained model. Understanding your impact across an organization ensures you can:

• Provide a business case for a promotion or job move
• Justify having your company sponsor a training course
• Talk about your work at conferences
• Interview well if you plan to move roles

11. Build Leadership Skills 

Best practice #12 will cover the different types of managerial career paths you may have in data science. But to get there, you will need to build leadership skills as an individual contributor. Leadership skills such as delegation, prioritization, giving feedback, and setting a clear direction are essential to build. Even if you do not manage people, you can still learn and apply these skills when you’re chosen to lead a project as an individual contributor.

There will be many opportunities to build these skills throughout your career, but this may also require you to leave your comfort zone. Moving out of your comfort zone and into the stretch zone is where our best learning and successes happen; don’t shy away from these opportunities as they represent excellent chances for growth which will pay dividends throughout your career.

12.  Choose A Career Path 

Career paths aren’t always straight lines nowadays. You may move laterally to a different team or project or move in and out of management roles. Meta has categorized some of their career paths here, stating that “...becoming a manager is not a promotion. It’s just another way to get work done”.  This philosophy is common across organizations, particularly those that are flatter in structure, with many teams rallied around projects rather than job functions. For Meta, the core career paths are

• Senior contributor: While senior contributors don’t generally manage people, they do often mentor junior team members and can have a significant influence on strategy and decisions. This career path involves mostly hands-on work and coding.
• Tech lead: This career path is a split between contributing to technical work and managing a small team of people, often other data workers.
• People managers: This is the traditional management path where you do very little coding and instead have high-level meetings and practice day-to-day people management.

It’s good to understand where you might want to end up later in your career. If people management isn’t for you at the moment, you can invest in your technical skills more, though if the opposite is true you may want to network and scout for opportunities to learn leadership skills.

Bonus—There Are Many Ways to Succeed in Data Science

This article has plenty of advice for data scientists within an organization, but there are many ways to succeed in data careers. Academia, teaching, freelancing, and content creation are all ways you can make the world a better place, learn or develop something, or increase your earning potential. Don’t shoehorn yourself into a single company or career path—consider what success looks like for you and go out and execute on your vision. 

More resources for you:

• Subscribe to the DataFramed Podcast
• Check out our certifications
• Watch this webinar on building a data science portfolio