What to Expect From Your Career Path as a Data Scientist

 

Data science is an exciting field. The possibilities within it are expanding, both in terms of applications and job opportunities. Whether you’re considering entering the field, already have your first data science job, or you’re already a few years in, here’s an overview of what to expect when you’re expecting to stay in data science.

Like with most career paths, the skills needed at higher levels build on those developed at the lower levels. It’s a good idea to look at the skills needed and the tasks you’ll be expected to handle at the level above where you are now. By looking for opportunities within your work to develop those skills, you can demonstrate your value when you go for your next promotion.

There’s a decent amount of overlap between positions that are directly above or below one another. When I worked in data science, my more senior coworker’s job wasn’t too different from mine, but they handled much more complex projects. Our boss was on another level, and the higher up the management chain you go, the less likely it is that you’ll be expected to or that you’ll have the opportunity to do technical work, like implementing models. This pattern holds true for a lot of the tech world, and data science is no exception.

As long as you’re still an individual contributor (IC), you can expect to still deal directly with data and models. Once you start managing others though, get ready to swap hats. You’ll have to interact a lot more with people - both the people who report to you and partner teams or customers. A big part of data science is turning business goals or requirements into data-driven insights, so you have to talk to different teams on the business side to understand the problem.

Each company has its own way of drawing lines between the levels. This is a rough guide based on an aggregation of a lot of different data science career patterns at diverse companies, but it could be different at the company where you work or where you’re applying.

Let’s dig into the different steps on a data scientist’s career path.

 

1. Data Science Associate

 
For this position, you’ve either just graduated from university or you’re just getting into data science. Most data science associate positions require minimal experience, but you’ll have a leg up if you’ve managed to gain practical experience with internships or side projects.

 
Skills Needed to Get the Job

A data science associate must be able to identify data sources and combine or aggregate data. You might be doing this kind of grunt work for your more senior colleagues. You will also be expected to perform data preprocessing. Some data science companies might expect you to come up with processes to clean, integrate, and evaluate large data sets from diverse sources.

Additionally, a data science associate will be expected to develop predictive models and present their findings with data visualizations. You should be quite comfortable with statistics and machine learning, both in theory and practice.

An important part of starting as a data scientist is being familiar with the context area. If it’s a plain tech company, this may not be necessary, but if you are analyzing data on the wear and tear of roads, you should learn as much as you can about physical infrastructure to better understand the data and your findings.

 
Projects You Will Likely be Tasked With

As a data science associate, your company may ask you to do a variety of basic data science tasks. These might include mining and analyzing existing data in the databases. You might also implement algorithms to maximize data extraction or design a tool to track and analyze the performance of a project or model.

Your main task will be organizing and using predictive models to produce insights.

 
Goals of the Position

To master this position, focus on learning the tools and techniques of the trade. You should master every kind of model you get to develop. You should also get to know your company’s tool stack, as a data scientist that knows how to use the tools around them is much more productive than one that does not.

The good news is that you’re at the start of the path. You can certainly expect support and guidance from your superiors, so make sure to ask for help when you need it.

 

2. Data Science Manager (IC)

 
A data science manager is still an individual contributor. The position requires between one to three years of prior experience.

 
Skills Needed to Get the Job

For this job, get ready to put on your talking mouth in addition to your thinking cap. You will need excellent communication skills since you will be working more with others in order to collaborate on more complex projects and gather requirements.

It’s critical that you are able to concisely and simply convey findings and make recommendations to non-technical people. You will have more interactions with people outside of your team. You will need to explain why your results are relative to their situation and how your findings should influence the solution.

On the technical side, you will be expected to implement predictive models. As a more senior member of the data science team, get ready to model automation initiatives, within the team and throughout the company as a whole. You’ll also need to drive technical innovation. Think of the tool stack your company uses and consider if there are elements that should be swapped out or added on.

You need to have significant experience building machine learning models of various kinds. You will be tasked with larger and more varied projects, and you’ll need to decide which models will work best for the situation with less supervision from your boss or team members.

 
Projects You Will Likely be Tasked With

As a data science manager, you might partner with the marketing team to come up with experiments. Data science managers also present findings to upper-level management, either in the form of a report or presentation.

You might also validate test data for representation, bias, etc. In addition to implementing models, you might be asked to develop solution prototypes if you’re working on a data science product. Even if the product isn’t focused on data science, you might still internally productize predictive pipelines.

 
Goals of the Position

The goals of this position are all focused on expanding the skills you developed as a data science associate and developing new ones to expand your impact. You will need to understand how to use the tools and work with the data. You should also improve your stakeholder management skills in terms of understanding their goals, concerns, and confusion. Be ready to proactively address their questions and provide them with solutions.

You may still need some help, but you should become more and more independent.

 

3. Data Science Senior Manager (IC)

 
A data science senior manager should have around four to five years of experience. At this stage, you are still an IC and do not yet have any direct reports.

 
Skills Needed to Get the Job

A data science senior manager should be able to leverage wide array machine learning models and statistical techniques. You will need to find and evaluate novel data sets.

Get ready to bring insights out of messy data, such as data from multiple languages, unlabeled, etc. You will also be expected to lead the deployment and maintenance of machine pipelines in production environments, which means you have mastered the dev ops of data science and can execute this task with masterful skill.

 
Projects You Will Likely be Tasked With

As a data science senior manager, you should provide thought leadership on analytical frameworks, code, data sharing, and more. You will need to serve as a kind of technical data science guru for your team and partner teams.

You will also develop and maintain close relationships with research, marketing, and product teams in order to align analytical efforts to business goals.

You will of course apply in-depth knowledge of deep learning, supervised learning, and/or unsupervised learning, depending on what is needed, to complex data sets. You will also need to collaborate with data engineers and platform architects to “implement robust production real-time and batch decisioning solutions,” according to some positions open at Apple.

 
Goals of the Position

A senior data science manager is really quite completely independent. You will be providing more junior members on your team with technical advice as well as help them navigate scenarios with partner teams.

You should be able to manage stakeholders, their expectations, requirements, timelines, and problems. You will need to manage entire projects with minimal supervision. The projects you are tasked with will grow to be more complex and might well lack clear-cut solutions.

 

4a. Staff/Principal Data Science Engineer (IC)

 
A staff or principal data science engineer has five to ten years of experience. This role is that of an individual contributor, so although you are very experienced and a true data science expert, you aren’t managing other people.

 
Skills Needed to Get the Job

This isn't your first rodeo. A principal data scientist can turn vague ideas or hunches into crystallized, numerical insights that drive improvements and profits for the company. You are an absolute whiz when it comes to implementing models, building pipelines, and analyzing data.

You have experience turning local models into live data science pipelines that expand the functionality and value your models bring to multiple teams or product groups.

Despite not having any technical reports, you are eager to mentor more junior members of your team.

 
Projects You Will Likely be Tasked With

As a principal data scientist, you are considered to be one of the most technically skilled data scientists around. You will be given some of the most complex problems and will be expected to run experimentation initiatives, build and maintain various pipelines used for model deployment and continuous operations.

You should deliver more accurate, leaner pipelines that are better able to be productized. Your data extraction skills are unparalleled, and they allow you to garner more insights from data than more junior data scientists would be able to produce.

 
Goals of the Position

As a principal data science engineer, you take on complex projects without any supervision. You own the product roadmap. Principal data science engineers find opportunities to enhance their team’s or department’s capabilities by applying data science principles to uncovered areas or championing the art of automation.

Your main job is to influence the solutions and their roadmap. You will have many different stakeholders across many different departments, so be prepared to elegantly handle their competing priorities and requirements.

You may not code a lot these days but you still review a lot of your team’s code.

 

4b. Data Science Director/Group Manager

 
A data science director or group manager also has about five to ten years of experience, but unlike all of the other roles listed until now, a group manager can expect to have about 4 reports. The exact size of the team that reports to you will depend a lot on your department and the company where you work.

 
Skills Needed to Get the Job

As a data science director, you’ll need broad knowledge and experience in most areas of data science. You should be familiar with Big Data tools (Spark, Hive, etc).

You should also be able to “comprehend and debug complex systems integrations spanning toolchains and teams” as a data science director. You can expertly identify stories behind the patterns in the data and distill analytical insights into concise, business-focused takeaways.

Your whole team looks up to you, so your skills and knowledge should allow you to be an all-around technical expert and support your team.

 
Projects You Will Likely be Tasked With

Data science directors collaborate with business teams to identify opportunities, step through requirements, and come back with technical solutions from their team. You must attract and mentor top talent, as the job market for data scientists is competitive if you’re the employer.

You will need to define and optimize data science strategy at a department-wide level. This involves a lot of different aspects, so you’ll need to have a big-picture idea that you can follow through with on a detailed level.

You may be expected to bring (financial) benefits to the company by proving the benefits the company has enjoyed by following data science recommendations. Be ready to justify the worth of your team or explain why it needs to expand.

 
Goals of the Position

You have direct reports who you manage and develop. Your job is to coach your direct reports and help them complete their job. You will need to guide your reports towards a solution to their problem without actually solving the problem. You essentially need to teach them how to do their job without doing it for them.

You need to resource the work against your team properly, meaning that you help prioritize the work and assign the work to a member of your team. You will help fill gaps when someone is on vacation and properly track and resource all of the work in order to fulfill the expectations and meet the timelines of partner teams.

 

5. Senior Director/VP of Data Science

 
At this point in your career, you have 10 or more years of experience. You will be managing an entire department or multiple departments.

 
Skills Needed to Get the Job

You need to have multiple years of experience managing others. This experience should have allowed you to hone your skill for deep technical coaching and mentorship.

You need to have extensive personal experience with data science models and have experience with building scalable data pipelines.

As the head of a department, you will need to thrive in a cross-functional, collaborative setting. You also need to keep customer needs and behavior at the forefront of your mind as you make decisions that will affect entire product lines or company-wide strategy.

 
Projects You Will Likely be Tasked With

You will serve as a subject-matter expert on modeling and data science in order to technically coach other data scientists. As a general champion for data science within the company, you will need to work with partner engineering teams to drive the integration of data science and automated solutions into products throughout the company. Wayfair expects their senior directors of data science to “understand, frame, and translate complex business problems to analytical problems that are well suited to machine learning solutions.”

Senior directors of data science are responsible for the roadmap, planning, and delivery of teams’ and departments’ whole portfolio of machine learning products. You will need to collaborate with product leads across multiple departments to ensure that the capitalization of data science solutions and benefits is maximized across the company.

As a member of the senior leadership team, you must bring a data science perspective to the boardroom to help drive the success of the business. You should bring data science to every corner of the company that could benefit from it.

 
Goals of the Position

As a senior director, you are a manager of managers. You have to have a vision of where the department needs to be and you guide the managers who report to you to execute against that vision. You will be more focused on the overall performance of the company and should frequently consider how your team can help contribute to the success of the company.

 

Final Thoughts on the Data Science Career Path

 
It is important that you always keep an eye out for new areas to apply data science (needed for senior roles, a great way to highlight yourself in more junior roles) at your company. You’ll probably always be able to keep coding if you want to, so consider whether you’d like to rise through the ranks as a manager or as an individual contributor. Data science teams are usually small, unless they are working on a data science product, so you don’t have to spend all of your time managing.

It is generally good to keep an eye on the “cutting edge” of data science. Keep yourself relevant and consider whether new technologies, tools, or solutions could benefit your team. Data scientists are hard to come by, so if your career isn’t progressing as quickly as you’d like it to or you’re not getting to work on the kind of projects you want, either explain it to your boss or apply elsewhere.

Keep your focus on the position above yours and hunt down opportunities to develop and showcase your data scientist skills.

 
 
Nate Rosidi is a data scientist and in product strategy. He's also an adjunct professor teaching analytics, and is the founder of StrataScratch, a platform helping data scientists prepare for their interviews with real interview questions from top companies. Connect with him on Twitter: StrataScratch or LinkedIn.