These days, it seems like everyone wants to be a data scientist. In fact, more and more universities are attracting and producing data science graduates—
You may recall
According to the 2017 Kaggle Data Science Survey, 37 percent of those who work with data report that a lack of management or financial support is a significant barrier to their work. This management barrier is only topped by lack of data science talent and dirty data, so more than likely you’re looking for both talent and managers.
Now that we’ve cleared that up, let’s consider the skills an organization needs to identify when hiring the individual contributor or manager.
Individual Contributor
There are three types of data science individual contributors:
- Data analyst
- Machine learning engineer
- Research data scientist
The first step is to understand who you are hiring and what you want them to do.
The data analyst role is suited to most businesses. Able to convert business challenges into opportunities for data analysis, the analyst often bridges the gap between technical and practical.
A machine learning engineer is looking to make an algorithm run quickly and in a distributed environment. Asking them to analyze data and find nuggets of relevant business insights isn’t their forte, but an ML engineer can select the appropriate algorithm and implement it within the company’s production system without introducing a bottleneck.
A research data scientist is interested in investigating cutting-edge techniques or inventing new techniques. This role usually requires a Ph.D. Extreme familiarity with the underlying mathematics is a must. It’s important to note this type of individual contributor would be bored out of their mind working on everyday-business problems.
Manager
The manager is the ultimate bridge between various technical roles, business stakeholders, and other leadership. Managers are frequently facilitating their teams’ best work while ensuring outcomes are mapped to business goals and prove ROI.
Breakdown of responsibilities of each role:
| Analyst | ML Engineer | Research Scientist | Manager |
Cleans, massages and organizes data (big and small). Performs descriptive statistics and analysis to develop insights. Builds models and solves business needs. Focuses on story-telling and visualization. | Develops, constructs, tests and maintains architectures. Works on databases and large-scale processing systems. Focuses on automating a prediction/ML algorithm to work in milliseconds and on large data. | Creates new algorithms. Researches topics of Machine Learning. Always playing and investigating new platforms. Publishes research papers in journals.
| Manages a team of analysts and data scientists. Strong leadership, project management, and interpersonal communication skills. Connects analysis to business outcomes. Creates room for the data science team to operate without undue influence. Understands database systems (SQL/No-SQL), predictive and prescriptive analytics. |
Mentality I will find a business solution. | Mentality I will make the business solution run faster. | Mentality I will create a new algorithm. | Mentality I will allow my team to create a solution. |
As you can imagine, these responsibilities require various skills to varying degrees. Illustrated below are some of the skills needed for each role, taken from a
With these skills and responsibilities in mind, you’ll know what kind of talent to look for in 2019, whether you’re hiring data scientists or their managers.
