Quick Summary - What kind of data science teams are there? What are the roles and how does everyone fit in a data science team?
Lastly, we take a quick look at the advantages of working with distributed data science teams by sourcing with Ukrainian IT professionals.
7 mins read
April 12, 2021
Traditionally, businesses have used historical data and metrics to drive a wide range of business decisions. Year-over-year growth provides guidance on trends to help set future sales targets. Demographical data helps to define target markets and set prices. Historical season and holiday sales performance helps to determine promotion strategies and prices. So, what’s the problem? A lack of accuracy, lack of granular data, new players and new products entering the market, and a lot more. As they say, “Past performance is not indicative of future results.”
Big Data proposes the ability to see what is happening today so you can make data-driven decisions in real-time. The more data you have and understand, the more profitable your decisions are likely to be.
There is a global shortage of software developers in general, but the need for data scientists and ML/AI specialists is particularly acute. As a result, companies have to tailor their teams according to the talent they have or expect to bring aboard. Whether working in a centralized or distributed environment, you’ll be looking at three main types of data science teams:
Data science teams are often structured and composed similarly to Agile software development teams. The way they are managed is a bit different, in that the work of a data scientist isn’t something that always translates well to tightly controlled sprints. Whether you’re resources allow for a software development team, hybrid team, or specialized data science team will largely determine the roles you can fill and optimal team structure.
Below, we cover many of the roles typical to a full-fledged specialized data science team. Typically, it will include one project manager, data science architect, scrum master, and business/data analyst. However, it may include one to several data scientists, data science developers/engineers, and subject matter experts. Though beyond our present scope, many companies have also added Chief Data Officer (CDO) to their C-level roles.
Wages are drawn from the US Bureau of Labor Statistics (BLS) and Payscale (P).
We welcome you to compare wages with our calculator. Select the roles you need to fill, how many, the term of the project, and email address and our cost report will be on its way to you.
Ideally, this individual should be able to define and prioritize project requirements from a business and technical perspectives. It’s possible for the two roles to be divided as is often the case when a non-tech company contracts with a software developer. The client’s product owner defines the business objectives for a product manager who details the technical requirements to make them happen. It’s also their responsibility to verify that the product that’s delivered meets the business and technical requirements.
A team may have just one to several data scientists so their range of responsibilities can vary. While specializing in analytical and quantitative skills, they also need solid programming skills and knowledge of your business/industry. A data scientist must be able to handle the entire lifecycle of data. This includes its acquisition, cleaning, validation, and analysis for the creation of models and algorithms to ultimately generate insights to help your business or solve a problem. It’s likely they will also have a hand in the creation of self-service reports so other data stakeholders can more easily analyze data for valuable insights. Their actual focus depends in large part upon the ML/AI and data science skills of the overall team.
are involved with the design of data collection, how it is stored, the processes of analyzing it, and making sure it fits with your business requirements. They’re dealing with the volumes amounting to Gigabytes and Terabytes, potentially on a daily basis, spanning potentially a hundred or more different file formats. Facebook deals with Petabytes on an hourly basis. They need to balance the ease and cost of storage with the ability for stakeholders to rapidly query it. Their main requirement is having a very strong technical knowledge.
Teams may have several developers with responsibilities divided according to their expertise. Their focus is developing Big Data, neutral networks and/or advanced analytics applications. This entails developing the user interface, features, backend, and defining data attributes to maintain consistency in their datasets. They will also be involved in the capturing and cleaning of Big Data, making it ready for your data scientists.
Teams may have several engineers/developers with specialized technical skills also strong in mathematical/analytical problem solving. It’s not necessary for engineers to necessarily know the industry or business as long as the Project Manager and SMEs who do are actively engaged. Actual job titles and responsibilities can vary by team.
Business Analysts (BA) and/or Data Analysts (DA) are the “bridge” between your business/industry and all things Tech. In the data science team, they’re a jack-of-all trades relative to industry, coding, and analytics. They are responsible for gathering, validating and organizing statistics in ways your team can use. Those with a more BA bent are also involved with system and process improvement to help solve business problems.
focuses on keeping the team in the process. They assign tasks in your project management software like JIRA, Wrike, Trello, etc.; guide your daily standups, retrospectives, and help clear any obstacles team members are encountering. They don’t need to know the industry or how to code. But, as is always the case with Agile development, cross-functionality in the team is an advantage, so it’s possible for this role to be shared possibly with the PM, DA/BA, or an SME.
Subject Matters Experts or SMEs only need specialized knowledge of your industry, and ideally, your company/brand history. They know how to actually use the data your data science team generates. They are responsible for guiding your data science team on how the data can help generate actionable insights.
The SME role can be drawn from potentially anyone with significant experience in your company/industry, even if they don’t know the difference between RAM and ROM. But, it’s also possible for any/every member of your team to be an SME. Given the shortage of qualified IT professionals, it’s simply much harder to find engineers and data scientists with industry, say nothing of company, specific knowledge.
These days, distributed development teams of all types seem to be the norm with so many areas still in lockdown. There’s not a lot of difference in the dynamics of managing a data science team in a distributed environment versus other software development teams. One difference is the potential savings you can find on developer wages compared to the fully-loaded costs of in-house developers and data scientists.
If your business or startup needs to ramp up its data science team rapidly – we can help you do so on an organized team-wide basis. As we source our IT specialists from Ukraine, you’ll be able to work with some of the most technical and analytics savvy engineers available, worldwide. According to SkillValue, Ukrainian developers ranked #5 globally with a technical rating of 93.17% compared to 84.59% for US-based developers. All of the Ukrainian IT staff we recommend to our clients also have a high degree of fluency in English (high-intermediate to advanced).
But, as is the case with data scientists everywhere, it can be had to source individuals with specific industry expertise. For many roles on a data science team, industry expertise is not essential. However, if you are looking for a long-term data science solution, you might start with ML/AI developers and specialists and pair them with your subject matter experts. Developers are very bright, and few things are more complex than ML/AI. They can come up to speed on the specifics of your industry much faster than a non-technical SME will ever be able to acquire ML/AI skills.
