In the past few months we’ve made a lot of progress reorganizing research and development in Zalando. We recently launched Zalando Research as a place where we can focus and conduct cutting edge research, as well as contribute actively to the research community in the areas of machine learning (ML), AI, natural language processing, and deep learning. We’ve also started organizing our teams to make sure we have a good ratio between research and engineering for every delivery unit.
We have created three different job roles and profiles that outline how we organize research within Zalando. All three roles have one thing in common: Each person is a researcher. What makes them different is their focus and interest. We have:
As we have iterated and refined these roles for the purposes of our newly launched lab, I wanted to share briefly how we distinguish them.
Firstly, research scientists are very senior and experienced researchers, have a PhD in Computer Science or related discipline, a minimum of 3-5 years of relevant research experience (either in academia or industry) in the area of ML, and a strong publication track record.
Research scientists are not engineers, and they work at Zalando Research to produce world class research, to help shape our products, publish and present papers at top conferences, file patents, and otherwise help to promote Zalando Research as as one of the premier research labs in the world. A research scientist usually focuses on continuous learning and experimentation as part of their research activity. We also offer a place for research scientist postdocs who have recently completed an excellent PhD, in order to provide an environment that helps them strengthen and grow their research experience.
A data scientist is also a researcher, but during their education focused primarily on, and have a very strong background in, data science. The area of data science typically comprises topics such as statistics, applied mathematics, operational research, data mining or modelling, and related disciplines.
Data scientists work in the context of a delivery team, and can do basic engineering tasks to help them complete their experiments, but are not aspiring to become engineering experts. Similar to research scientists, data scientists are encouraged to publish papers, file patents, and actively contribute to the research community, as long as it is in line with the delivery goals of their team.
Last but not least, the research engineer combines strong research skills with comprehensive engineering experience. Similar to full-stack engineers who combine frontend and backend expertise, a research engineer combines research and engineering: They are able to work relatively independently on researching complex ideas and pushing them into production, by writing production-quality code.
Usually, research engineers work in different “modes”. They start out with an idea, explore and evaluate it (“research mode”) for some weeks, and then once they have proven the merit of the idea, they go into “engineering mode” to implement their idea into a production system. This process is then repeated for new ideas and projects.
Often research engineers get stuck in maintaining a system that they had previously developed, which prevents them working on more original research. Therefore we recommend that research engineers strive for an ongoing balance between research and engineering. Just like research and data scientists, research engineers are encouraged to publish papers, file patents, and actively contribute to the research community, as long as it is in line with the delivery goals of their team.
Why do we feel that it is important to know what research role you are currently fulfilling? We want all of our research colleagues to be able to carve out a career within the industry.
As a research scientist, you may feel the urge to get closer to product and delivery, therefore migrating more into a role of a data scientist or research engineer. In the latter, ramping up on engineering skills is important. If you wish to remain a research scientist, continue broadening and deepening your research expertise to grow your career.
As a data scientist, you may want to improve your engineering skills to become a research engineer. Or perhaps you would like to focus more on core research, not within the context of a delivery team or product, to eventually become a research scientist. In this case, you would need to work on your publication track record, for example.
As a research engineer, broadening either your research or engineering skills would be appealing, or you might be aiming to work in a concentrated research environment. To realize this, you would need to work on your publication track record or on fulfilling all the requirements needed to become a research scientist, listed above.
I hope this article helps to clarify the differences between the research related job roles we offer at Zalando. If you’re interested in these job opportunities, more information can be found here.