In this episode, Jason talks to Natalia Lyarskaya, Chief Data Officer at ZestMoney, and one of the DataIQ top 100 most influential people in data. Their conversation focuses on building data science teams, how data is measured, and how to apply a business and product outcome-focused approach to it.
Listen to this episode on Spotify, iTunes, and Stitcher. You can also catch up on the previous episodes of the Hub & Spoken podcast when you subscribe.
For more on data, take a look at the webinars and events that we have lined up for you.
The key to building a successful team is really for the business to understand why they need a data science team.
[00:55] Natalia shares her role and experience as a Chief Data Officer
[03:06] Natalia talks about her set of principles in building out a data science team.
[07:22] Is there a typical type of team to build early on versus later?
[13:29] The best way to set up a data science team
[17:31] Building a product that is fueled by data and AI learning machine
[20:25] What is the core skill set for a well-rounded science function?
[24:09] Natalia shares their guard rails in managing diverse teams.
[27:15] Different cultural challenges in setting up a team from different countries.
[30:23] How has data science changed, and where is it going in the future?
Data-centered businesses make decisions based on the data, instead of gut-feeling or assumptions. Consider that the key product is built by data. Data is the enabler of the measurements and how the product needs to be treated. It is more about how the users use the product and this is what helps to see and define what is the next step, or what improvements can be made to the product.
There is no common approach to building a team. It’s a station where you need to consider each case separately, depending on the size of the team, depending on all the purposes, and why the business needs the team.
A key principle in building a team is the need to align the mission, and the vision, for the data science team with the vision of the business or the company in general. This is what defines to a large extent, the success of the data science team as well.
Looking at the coding abilities of the candidates alone is not the most effective way to build a well-rounded team. It is crucial to see that every new person coming to the team actually has an additional new skill set to bring in with them. Look at people that come from completely different industries, for example from retailers, e-commerce, delivery, or whatever other industry you might think of. It gives a nice breadth and diversity of thinking, along with different angles to a problem.
There are a lot of cultural differences in specifics. Not only to the end consumers and customers, but even when hiring people and working in the team. You may find that in certain markets or geographies, technical tests need to be conducted to gauge qualification of applicants for the data science roles. Whereas, say in the U.K. market, more data scientists come in with a strong background in Ph.D. in computer science. They are very much research-oriented and come with necessary technical abilities, and thus have strong profiles.
Changes in data science have accelerated in the past 10 years. People have moved away from the hype of A.I. (big data). A lot of people in the industry realise that it’s no more just a buzzword. But actually, it is the internal I.P. of the company and the way they build a competitive science data team as an advantage. If we don’t move ahead of just buzzwords, the product and the businesses won’t survive.