Data science is a career area that has exploded in the last years and offers many opportunities for scientists whose projects involved the analysis of large datasets. We recently spoke to Michael Kachala, an EMBL alumnus who is now Global Head of Data Science at Bayer Consumer Health, about his career path and suggestions for aspiring data scientists. He advises that data science is a great career for people who love to learn new things, and recommends taking up pet projects and Kaggle competitions to demonstrate your acquired skills. In addition to technical skills, Michael reports that communications skills are key to success in this field.

You also can learn more about this career area at:

• From doctorate to data science: a very short guide – NatureJobs
• So you want to be a data scientist – Nature Jobs
• Kaggle – a community of data scientists and machine learners

You were a predoc at EMBL Hamburg and you are currently Global Head of Data Science at Bayer Consumer Health. Could you explain how your career developed?

The first step was actually the hardest when I was switching from academia to the business world. It took me several months to find a suitable position. For our readers: don’t be too disappointed if you don’t find the perfect position immediately after your graduation. It was really a job in itself. The more interviews you attend the better you get. Also, attending interviews helps you to understand the different aspects of different jobs.

I ended up doing data science but I was also interviewing for other positions in i.e. translational science. In the end, I realized that data science is what I really wanted to do and I started my data science position at Bayer, a large pharmaceutical, consumer health and crop science corporation.

During the next four years at the company, I was working on various projects covering different functions from R&D to Marketing and Sales. This helped me a lot in understanding the business and to develop my career. When the new position was opened for Global Head of Data Science in Consumer Health, I already knew a lot about the consumer health business. And I already knew how data science can be applied to different aspects of this business. So, for me, it was a reasonable transition to this position.

What does your current role as Global Head of Data Science for Consumer Health involve?

It involves a lot of different things and it is quite a new group – we only exist since this year so it might change. But for now, my top three priorities are project delivery, people coaching and development of the data science strategy.

Project delivery is probably the most obvious one. I need to assure that the projects that are delivered by our team members are of high quality, that the development is done in a fast and agile manner and that the final result meets or exceeds expectations of our business stakeholders.

Of course, coaching people is a very important part. Data science is still a rapidly developing field and you do not get fully trained data scientists from the university or even other companies. They really need to learn how to apply data science to the business problems that we have in consumer health. And also, to use the technology stack which is common in our company.

Finally, data science strategy. Again, it is a new field and there is not a common understanding in the business what it exactly means and what kind of benefits data science can bring to the organisation. Thus, a large part of my role is also education and formulation of strategic projects that can help the business.

At EMBL Hamburg you were working on data analysis of structural bioinformatics and SAXS data. How difficult was it to convince industry that your experience was relevant for data science roles?

It was actually a bit easier back in 2016 because there was much less awareness about data science. And the work that I was doing in Hamburg around methods of data analysis was actually quite similar to the data science projects which I had at Bayer. I demonstrated that some of the skills which I acquired in programming languages, also in the algorithms for optimization, were also very useful in a business environment. Because in the end it is the same types of data. Just in different projects.

Do you have any tips for scientists now who are trying to translate their academic bioinformatics or computational experience for industry roles?

The situation has changed significantly in these years but also in a good way for data scientists because there is much more material available. Many more online courses which you can take. Also, many more opportunities to do your pet projects using Google Collab. And my strong recommendation is to do some pet projects. Small pet projects can teach you a lot and it is easy to find something interesting for you such as something related to a hobby of yours. Some people do it for their favourite sports. You can team-up with other aspiring data scientists, do something interesting for yourself, like a hobby and also participate in Kaggle competitions which are more popular right now compared to 5 years ago and you really learn a lot from these competitions, especially working in a team.

What are the main challenges of working in data science roles in industry?

It depends a lot on the company and the area where you are working. In digital-native companies such as Google and Facebook, there are completely different challenges. But for us in a traditional business, the biggest challenge is getting the data. Because it is somehow scattered across the company. So, if you want to build a nice project you need to connect many dots to get all this information. And I think it will stay like this for the next 5 years until we have really well-structured data sources inside the companies.

The second challenge is to incorporate the data science solution into the existing workflow because people are used to doing things like they did for many years. Now we come with the sophisticated data science or AI solution which completely changes their business processes and you need to invest some time into convincing people to adopt these solutions.

What do you find the most enjoyable?

The most enjoyable aspect is when after we convinced them that it is a nice solution and we make it work, which can take some time and effort, we see that our solution actually makes life easier and better for our end users. It can be internal employees who now save time on their day-to-day duties or consumers which now have access to better everyday health.

What skills and character traits do you think were needed for you and other data scientists to be successful in such roles?

As the field is very dynamic, you need to be ready to learn a lot. Almost every year there is a completely new solution in the field you are working in and you need to be ready to study, to learn technologies, to learn new frameworks and sometimes you don’t even have the proper documentation. Then you have to follow some YouTube tutorials or follow the people who actually developed this software and read their personal blogs. So, it is also up to you to find information if you want to be successful. But the community support is great. If you ask a question about a new technology people are really happy to answer it.

The second important skill is communication and this works both ways. On the one hand, you should be able to deliver your ideas in a nice way. Either by public speaking or by a presentation. On the other hand, you should be a good listener because if you don’t listen you cannot understand the problem the business is having and what is important for them.

These are the two key soft skills to be successful.

In your current role, you are involved in hiring. While hiring, what are you looking for besides those skills?

To be honest, I assume that when someone is already working as a data scientist, they are eager to learn and that they understand that there are a lot of things they need to learn. The communication skills we look at during the interview.

The key factors for me is the most recent experience. The last two-to-three years of experience and what kind of projects this person delivered. What kind of technology was used because as I said it is a dynamic field and so the relevant technology is only two to three years old.

Of course, fundamental knowledge is very important. You should have a good background in mathematics and ideally in computer science but this is not decisive. It is not a software engineer position so you don’t need to be excellent in coding, but a good knowledge of R or Python will help you to create better data science solutions.

Regarding skills, I like to see in CVs where people state their skills clearly such as the programming languages they are doing now, the technologies, frameworks and libraries that they use in their work.

How much weight do you place on courses on data science from Coursera or somewhere else versus hands-on-experience?

This is a tricky question. Obviously, it is always helpful to have hands-on-experience. It is the best way to learn. But there are two big caveats of hands-on-experience which I have seen in many CVs and in my personal career: that first of all you don’t have that exciting projects in real life compared to an academic environment or during your courses because we don’t use the most recent technologies in our projects. We tend to use more robust solutions.

Also, if you are already an expert in something you tend to use the same technologies, the same tools over and over again. I have seen CVs of people where they spend two years in a company and they do the same thing over and over again. Which is nice if you want to become an expert but you don’t learn new things. This is why I usually consider Coursera and other online course providers on a CV as a sign that people are eager to learn and as a justification for the skills they put in their CVs. If they say that, for example, they know TensorFlow and I see that TensorFlow course in their CV I tend to believe it.

In what ways is the work you are currently doing similar to your work in academia? In what ways is it different?

I wasn’t working as a manager in academia, so it is hard to tell what is the difference on that level. But the main experience is that the business world is much more dynamic. In academia, I can work several months or years on one topic and you want to get the perfect solution. In our world it is different. We have deadlines. In our company and practices, we usually have a three months project timeframe to deliver a prototype. This time pressure makes you choose not the most elaborate approach but a more simple solution that works well and that you can deliver fast.

Which I guess nicely links to the next question: There is quite a lot of time pressure. Do you have any tips on how to attain a good work-life balance in this successful career?

First of all, don’t work on weekends. Please let everybody who works with you know that you don’t work on weekends. That is definitively not a healthy habit.

To be honest, for me the greatest accelerator of finding a good work-life-balance was the birth of my daughter. When you have less spare time since you work much more efficiently. You start to prioritize ruthlessly and you cannot afford to stay 12 hours at work because you need (and want!) to go home and spend time with your daughter. It is not your time anymore, but time shared with your daughter.

Do you think there is anything we did not touch on that is really important?

First of all, talking about CVs: I would really recommend sticking to one page because nobody has time to go through even two-page CVs. If a person is interested in you they will go to your LinkedIn page and find all the necessary details there.

The second recommendation is a book which really helped me in understanding the pharmaceutical and biotechnology industries. It is called Career Opportunities in Biotechnology and Drug Development from Toby Freedman*. This was the most useful book for my career development.

And for the later stages of your career, I recommend the book Rise by Patty Azzarello, which contains advice on how to choose the right projects and establish communications to succeed in the corporate world.

Lastly, scientific writing is a very useful skill which is often overlooked. Anybody who defended a PhD thesis has written a very large chunk of text and this is a very important skill. And this is very easily translated to the business environment. So, you should mention it either in your CV or in your interview that you definitely have strong communication skills in writing as you have written a thesis or a scientific publication. But it, unfortunately, gets overlooked often to focus on other skills such as presentation skills.

So, would you say that writing skills are really important for your career?

Yes, it is getting more and more important. Especially when you advance in your career because you talk to people who don’t have time for two-hour-long meetings. They would like to read short memos. And it also helps you to structure your thoughts when you put it into writing and slides.

I think it is necessary for people to understand what kind of career opportunities they have, especially when talking about data science which is a new field and thus not very clear when it comes to how it works.

---

EMBL fellows and staff members can contact other former staff/fellows working in data science and other non-academic roles via the EMBL Alumni Directory. Please see our article on informational interviews for ideas on how to make the most of such opportunities. Michael can also be found on and twitter.

• Note for EMBL readers: the first book Michael referred to (Career Opportunities in Biotechnology and Drug Development from Toby Freedman) is in the EMBL Library and the careers service hosted an internal event with Toby in November 2020 a recording is available within EMBL in our career exploration resources (under ‘Recorded talks & alumni panels covering multiple career areas – full list’).