Data Science Career Summit: ‘Building a Data Science Career’ — Day 1
On November 18-19th AIgents and the European Leadership University organized the first Data Science Career Summit. The title of this first edition was ‘Building a Data Science Career’. In two days, 14 different talks were presented by Data Science consultants, team leads, entrepreneurs, recruiters and headhunters.
The program was organized for (aspiring) Data Science students, graduates, people in their early careers and more seasoned professionals that consider switching careers from another field. 300 attendees from 15+ different countries watched us live on Zoom where we covered topics such as:
- Why choose a Data Science career?
- Data Science job & career perspectives
- How to choose a program and institute?
- Understanding titles: ‘Who is who in the Data Science team’
- Skills: ‘What languages and tools to learn’
- What’s it like working as a Data Scientists (corporate vs entrepreneur)
Day 2 (for summary & highlights go here)
- Leadership skills for Data Scientists
- What’s it like working at a Machine Learning consultancy firm
- What recruiters look for in a resume
- How to prepare for your job interview
- Data Science case presentation
Here is a link to the full program and line-up: https://techminds.elu.nl/data-science-career
Meeting the Hosts
AIgents is a career community for Data Science and Machine Learning Engineers. It's used to discover career opportunities (like jobs and internships), events and training courses. You can join their Machine Learning Meetup or one of their LInkedIn groups to keep track of developments in your field of study. To connect, learn and collaborate with other students and professional go to 'Masters in AI' learning community.
Founded by Robbert Van Vlijmen, the community has become one of the largest tech communities in Europe.
European Leadership University (elu.nl)
The European Leadership University is a next-generation university focused on filling the skills gap that businesses currently face. Founded by Alper Utku, the university consists of programs based on the job market, investing in young professionals to give them the skills they need now.
Understanding Titles: Who is Who in the Team?
Vivienne Haring has a master's in econometrics and works as a Senior Data Science Consultant at Amsterdam Data Collective, which is a fast-growing consultancy firm focusing mainly on the financial, public, and health sectors.
Table of Content
- Type of roles
- Type of employers
- Job Titles
- Average Salaries
Type of roles
There are so many roles involved in data science that it can be confusing to understand who does what in a team. Ten years ago, data science was not well established in companies. Data Scientists had to be generalist professionals who would manage end-to-end data science projects. Nowadays, the data science pipeline is getting cut up into different pieces. As long as data is growing, the data science field is also growing, so more different roles are appearing.
In smaller companies, there are usually not many people in the data science team, so the Data Scientist is still an end-to-end position. In this role, the professional has to create datasets, ensure data quality, build analytical models, and keep close contact with the business stakeholders. On the other hand, big companies are adapting to more divisions in data teams, creating different roles, and even whole teams.
The data engineer is the one who really focuses on data architecture, doing the technical part by building the whole data warehouse. The data architecture designs how the data should be saved in the different systems, how the structure should be, and how the data should be linked to each other. The data strategist is business-focused, determining what should be the strategy about what we want from the data in the company and they can also request more data when needed. The analytics translator is the link between the business and the analysts, translating the models to the business stakeholders. The business analysts do basic analytics on the data, looking for trends. And finally, the data scientist, who is the link between the technical and analytical sides, performing an in-depth analysis of the data.
Type of employers
- Company (corporate): It can be large or small. The types of roles will vary according to the company size.
- Consultancy: It can also be large or small. The roles are divided depending on the size of the projects. For instance, in smaller projects, you can do an end-to-end role.
- Contracting (ZZP): You will be working on your own. This is indicated for more experienced professionals and the salary can be higher.
- Start-up: Most of the time there is only one data scientist. The roles in start-ups are often innovative and you have the freedom to do what you want. However, you don't have senior data scientists from whom you can learn.
The job titles will vary according to the years of experience and the type of employers. In general, we have the analyst position first, which is the start point after finishing your studies. After a few years, you can grow to a senior or manager position, a couple of years later you can be head of some department, and so on. Of course, as long as you grow the ladder you get more responsibilities.
The salaries in data science roles will differ depending on the job type, years of experience, and the country you live in. In the image below you can check the average salaries for different positions in the USA. In Europe, the average is salary is a bit lower but trends and the differences between the different roles are more or less the same. Check this article for salary trends in The Netherlands and EU.
Skills: What Coding Languages & Tools to Learn
Henriette Claus has a master's in econometrics and quantitative finance and has been working at Amsterdam Data Collective for two years as a Data Science Consultant.
Table of Content
- Overview of current tools
- Future trends
Current tools used for the different data teams
Over the years different programming languages took the lead among the most popular. Currently, the most popular programming language is Python, according to StackOverflow. However, it's important to note that the programming languages and tools required will differ from company to company.
- Analytics Translation: It uses mostly tools for visualizing data and making dashboards, such as PowerBI and Tableau. Many companies can still use Excel to visualize data as well.
- Business Analytics: It also uses tools for data visualization and for creating more interactive dashboards, such as PowerBI, Tableau, Python Dash, and R shiny.
- Data Science: Python is the most popular programming language in fast-growing companies, mainly for being all open-source. R is popular for medical research while some big companies, especially banks, prefer to use Matlab.
- Data Engineering: SQL is very common. Hadoop and Spark are popular for data storage and processing large amounts of data. For web development, we have Docker (an app for developers), Django (a high-level python web framework), Laravel (a custom software development tool), VueJS (also a great tool for building websites).
- Data Architecture: SQL is also very commonly used here, Java (for developing desktop and mobile applications),KNIME (a modular data pipelining concept, good for an agile way of working and for machine learning), Powerdesigner (used to visualize databases), and Azure (Microsoft cloud computing service for analytics and data storage).
- Data Strategy: For this team, there's no tool established, perhaps PowerPoint to translate the knowledge. In Data Strategy is necessary managerial skills.
Java is still one of the largest programming languages. All the big companies (Amazon, Netflix, LinkedIn) use frameworks based on Java.
Docker is a service to deliver software in packages called containers. Containers are isolated from one another and bundle their own software, libraries, and configuration files. They can also communicate with each other through well-defined channels. It's becoming the most popular way of working.
Although Python is already the most common programming language, it’s still gaining popularity. Python is all open-source which means that it's fast-growing and very easy to share code. Besides, we have a lot of opportunities and pre-written code for AI and machine learning purposes.
Swift is a programming language developed by Apple and is also completely open-source. It's a very intuitive language and it has grown very fast in the US.
DevOps Stands for development and operations, and the idea behind that is to work on very interactive feedback loops, called the eight loops. You can build your entire software or tool in small parts and go through this loop for each of those parts. It is generally very compatible with an agile or scrum way of working.
Don't get attached to just one language or tool, as they can vary from company to company. You have to be open to learning new tools and languages, in addition, if you already understand the basics and how things work, you can easily learn a new programming language.
Why Choose Data Science as a Career?
Longhow Lam is a very experienced data scientist who has been working with data for more than 20 years in different business areas. He discussed a few critical questions for people who are considering data science as a career.
What to expect from a career in data science?
Data science roles have changed many times over the years, from Applied Statistician to Data Miner, Data Scientist, Machine Learning Engineer, AI Specialist. To work in this field that evolves fast, you need to be prepared to be constantly learning.
Also, it's necessary to highlight that the core of data science is to solve problems. Many times, fancy projects such as deep learning models won’t solve the business problem. So as a data scientist, you should be willing to do things that are not always highly skilled data science techniques, but that will do what it needs to be done.
What is a must-have skill for data scientists?
The most fundamental skill for a data scientist is to be able to communicate with the business. As a data scientist, your main duty is to understand and investigate the business problem.
How is the day-to-day work?
Data science is a broad field and can vary from company to company. Sometimes, create a simple data analysis and some data visualizations is data science for a company. In other cases, data science can involve complex mathematical and machine learning models.
Nevertheless, any data science project will start with understanding the business, by talking to the owner of the problem. Then, data scientists will get the analysis and the models to investigate the problem, and finally, they will present the insights and ways to solve it.
What are the requirements to enter the data science field?
To enter in data science you should be eager to learn, as the tools and programming languages are always changing. You should enjoy maths and statistics that are the foundation of data science. In addition, as a data scientist you need to understand the business and know how to manage with the stakeholders.
How AutoML (Automated Machine Learning) is affecting the field?
AutoML allows you to quickly generate machine learning models which can save you time. However, creating a model is just a small part of the data science journey, the consulting part is more important and once you get a model, you need to investigate the data, deploy it, and explain to people how to use it. Therefore, AutoML will not replace data scientists but it can be used as a complement.
Why is ethics important in data science?
Being able to access and collect data does not mean that it is ethical to use that data. Data scientists may work with very sensitive data and it's crucial to understand what are the rights and ethics about it. We have to remember that we are talking about people and how we can influence them, so it takes a lot of responsability.
Tech Jobs and Career Perspectives
Teddy Dimitrova is very experienced in tech recruitment and talent management and has been helping startups and companies grow over the last five years, mostly in The Netherlands. She has discussed key marks for those looking for the first job or making a career switch to the tech industry.
Table of Content
- Education Vs. Skills
- How to get the first job
- Switching careers to Data Science
- Tips for interviews in The Netherlands
Education Vs. Skills
There are always many job openings in the tech industry and nowadays more companies are getting into data. However, various companies are just following the wave of data science without actually knowing what they are doing. This brings up some confusion about the job roles and what they are really looking for. Check this article to see more about the different roles in data science.
Many companies today don't have requirements based on education. Instead, they are looking for hands-on experience. Degrees are not so relevant in the industry anymore because they don't prepare young professionals with the required skills.
Currently, we can learn a lot without going to university. We have more practical education like MOOCs (Massive Open Online Courses) and Bootcamps that can give you the technical skills. Besides, the market is asking more for skills and experience, than for a certificate. Still, for those who want to follow a traditional education, a college degree is also highly recommended. You can learn a lot of theoretical and important things but keep in mind that you also have to build up your technical skills.
How to get the first job?
Traineeships are a great way to start because you will be learning a lot and they are very hands-on. About the employers, bigger companies and consultancies are often more open to investing in new talents and helping them to develop hard and soft skills.
Another way of catching the company’s attention is by participating in Kaggle competitions. That can say many things about candidates including that they are entrepreneurial, competitive, and want to achieve things. Besides, it’s a great way of learning and showing your technical skills.
It's hard to get the first job but once you get into the job market, many doors will start to open for you. Also, the sooner you start building a network with people in the field, the sooner you will find new opportunities.
Switching careers to Data Science
Companies care about people who can fix their problems. You can sell yourself greatly if you know how to translate your skills into things that matter for a specific position. So maybe you have transferable skills from your previous experiences such as good communication and presentation skills that could be useful for a position in data science.
Therefore, when applying for a job position, write down the skills you acquired in your previous experience that could be applied to the problem that the company is facing. Also, having great soft skills already gives you a lot of advantages in the field.
Tips for interviews in The Netherlands
The first tip is very simple, just be yourself. Be honest because only the companies that would really appreciate you will appreciate it when you’re honest. Also, be self-reflective, knowing your strengths and weaknesses, and what you want to achieve in the coming years is key for a successful interview.
Start Learning: How to Choose a Programme and Institute
Anya Tonne has a bachelor's degree in management science and mathematics and a master's in social and behavioral sciences. She works as a Consultant at Cmotions building predictive models, creating analysis, improving data quality, and optimizing data processes.
Table of Content
- What do you need to become a good data professional?
- Choosing a programme
- Choosing an institute
- Cmotions Academy
What do you need to become a good data professional?
Many courses are focused on the toolkit, like Python that is the most common programming language. Although they are important and you have to learn them, keep in mind that tools can change and you won't be using Python for the next 40 years.
- Understanding of the business process
You should understand where you can make a difference. What is very important about any data professional is being able to make a difference with the data. If you understand how the business operates you know where you can make that change and bring value with the data.
- Soft skills
Soft skills are important because you need to be able to communicate with your colleagues as well as to explain your findings. Besides, apart from coding, you should also be able to inspire colleagues to work with data.
Choosing a programme
It's more suitable to start by choosing the programme instead of the institute. To help with this decision think about the 3 questions below.
Do you already know exactly the kind of data professional you want to be?
If you do know it will be easier to choose. For instance, if you want to be a real data scientist, in the sense that you want to build predictive models, then you might be looking for a degree that has very rigorous mathematics and statistics in it.
Nonetheless, if that's not your main interest and you want to work more towards data management or data engineering, if you want to prepare the data, then maths and statistics might not be as useful for you. Even though it's still nice to understand those concepts, in this case, I would focus more on computer science and specific tooling.
Do you have any prior experience?
If you are switching careers you may already have filled a few gaps in the field. For instance, if you have a business background, then the business side of the education might be a bit less relevant for you.
Likewise, if you have research and analytical experience, maybe you have been working in academia, so you already have a lot of analytical projects, probably published papers, and those are all very valuable. In this case, you might don't need to do a full-on degree and just take a couple of courses.
Lenght of the programme and portfolio
It's also important to think about the length of your education. If you don't have a degree yet, it would be a good idea to choose a longer one to give yourself time to build your portfolio (Kaggle, GitHub). In university, you can also take research projects, like during a master's degree where you can dedicate a full year to your own projects.
Choosing an institute
The location is a good thing to check before choosing an institute. It's better to study in the same region where you want to work since you will start building connections there. The professors might be also doing projects in companies and will connect you with the right people. Besides, many universities have mentorship programs that can help you to connect with people in the industry.
In addition to classical education, many companies want to prepare young professionals who come without industry experience. For instance, at Cmotions Academy they have a six-week program where they cover the tooling that is mostly hot in the market and give real-life business projects to the candidates. Besides, they help with soft skills, presentation, and advisory skills. By the end of the program, they also help them to find jobs.
Portfolio Building and Job Searching as a Data Scientist
Violeta is a data scientist working at ABN Amro, a large bank in The Netherlands. She has a doctorate in applied econometrics and is experienced in machine learning, natural language processing, among other topics in data science. Violeta answered a few questions about her career journey and gave some valuable advice for aspiring data scientists.
Table of Content
- Skills you need as a data scientist
- Challenges when applying for jobs
- Overcoming these challenges
Skills you need as a data scientist
The image below shows an overview of the skills you might need to have as a data scientist. Some skills such as business and domain knowledge, you might acquire or improve only during the job. It's also important to highlight that in the data science field you will be constantly learning.
Challenges when applying for jobs
- The job market wants practical experience (projects) and not only theoretical. Therefore, it's important to have demonstrable programming experience/knowledge of open-source tools (preferably Python and SQL).
- You might find wrong descriptions of the roles in job descriptions. Often, companies oversold the job or what you will need to be doing in the job.
- If you are applying in The Netherlands and you're not Dutch, you will also have a language barrier. Although the Dutch language is not really necessary for the job it can be an issue, especially for small companies.
- Sometimes, even when it seems that the interview process is going well, they might just disappear. So be ready for a lot of ghosting.
Overcoming these challenges
- It's crucial to identify what skills you may be missing and invest in them.
- Connect the dots for recruiters and hiring managers explaining to them how your current training, educations, and experience are relevant for the position you're applying for.
- Invest in building up a network (attend MeetUps, conferences, events). If you get a recommendation from someone inside a company it will increase your chance of success.
- Build up a portfolio. Pick open-source data, if you have a preferred industry, then choose a problem from that domain. It's better if you focus on a problem you care about.
- Keep your GitHub up-to-date. If you don't have experience as a data scientist, then this is a way to differentiate yourself. Include not only data-related projects but also contributions to open-source packages, presentations, etc.
- Keep your Linkedin up-to-date so that recruiters can find you easily.
- Look for job positions on different websites. You can also work with recruiting agencies and depending on the company you might have a priority for an interview because they act as a first filter.
- Keep in mind that it will be hard to get the first job but once you get it will be much easier to find a second one. Besides, the more experienced you get, the more opportunities and choices you have.
Interview: What is it like working at ABN Amro?
Which job titles did you focus on during your job search?
— Mostly on data scientist and data analyst roles because 5 years ago the industry didn't have so much diversification in data science positions yet. Today, when looking for jobs in data science you should search for a variety of titles depending on what you want to do. It's also important to read carefully the job descriptions because the same titles might mean different things within companies.
What are you working on at ABN Amro?
— I'm working in a central role within the bank, serving various business lines almost like an internal consultant. It's very nice because you get to see the different parts of the bank function and we are always working on different projects. We also recently started on bigger topics that we think the data science community within the bank should focus on. For instance, I'm working with responsible machine learning including topics as explainability of machine learning models to ensure that our models are unbiased and don't discriminate against certain subgroups of the population.
How did you land your job?
— For my current job I was headhunted for a recruiter through LinkedIn which is quite common in The Netherlands, so always keep your profile up-to-date. I had a few conversations with the recruiter and hiring manager and after a few weeks, I followed the standard application process which is working on a case and giving a presentation for the rest of the team that I would work for. My first job was a bit more difficult I had to send several applications and got many rejections. To find the first job requires patience and persistence, also, it might not be your dream job but is a great opportunity to learn and grow to your dream job.
Would you recommend using a recruitment agency to find a job?
— Yes, I would. Not all job offers are on LinkedIn or other portals, besides, it can save you a lot of effort of checking on the various job posting sites. Also, when you work with recruiting agencies, you might have a priority for an interview because they act as a first filter. However, there is a large number of recruiting agencies and there is a difference in the quality of their services, so it's something that you need to try out and see what works best for you.
What is the bigger challenge for you in working in a large bank?
— One downside of working at such an organization is that is difficult to keep track of everything that's going on. Sometimes it's difficult to stay connected to all relevant stakeholders and with the various communities of data scientists and engineers. We have been working to improve that but at least a few years ago it could be the case that you start working on something but maybe a couple of years before someone else might have tried to solve that problem and there was a loss in translation.
What do you like the most about your job?
— The fact that we get to work with different business lines is very exciting because you're not always within the same domain. For instance, currently, I'm working with private banking, with our internal economic research group, among other things. I like about this diversity and the freedom we have during our projects to propose topics we think are important.
What are the most important skills for data scientists to do well at ABN?
— Well, not only for ABN but this would be in general to work as a data scientist. It's important to have a solid technical foundation as well as have good soft skills. Being a good communicator is essential to explain your solutions to stakeholders, especially for non-technical people because if they don't trust you're doing they won't use it. Also, it's important to be curious and stay motivated. Very often you might have an idea and when you try out it might not work, then you need to go back and come up with new approaches. Therefore, you have to be persistent but also realistic and pragmatic.
How long do you think someone should take to plan out a career switch to data science?
— How long the switch would take really depends on what you have been doing so far, what exactly you want to do in data science, and in some cases, you can even try to do the transition within your own company. In my case, I already had a foundation so it took me around six months to make a transition from my Ph.D. to data science, where I completed a data science specialization on Coursera.
Interview: What is it like working as a Data Scientist at Startups?
Martijn is an entrepreneur with a master's in data science and 5 years of experience in the field. He has worked in 6 startups over the last 4 years and currently has his own startup where he is working on a new data product for e-commerce companies.
How is it like to be an entrepreneur?
Be an entrepreneur is exciting because you have to work on all parts of a project, also, you have the freedom to be creative, come up with new ideas, and execute them immediately. However, you also have to be prepared to deal with the risks. It's essential to set up a financial buffer so you don't have to worry about money for a while. The good thing is that tech startups are very low-cost in general, so you don’t have to invest a lot of money in physical goods.
Differences between working in startups Vs. big companies
Overview of the main differences between startups and big companies.
Also, it's very relevant to know the data science hierarchy of needs if you’re going to work in a startup. The big difference between startups and big companies is that usually a startup still has the challenge of the lower stages of the hierarchy of needs.
Of course, large companies can also have this issue but usually, they already have a bunch of data that they either don’t know anything about it or don’t know what to do with it.
Besides, if you are a startup you need to create value relatively fast so you might not have 2 or 3 years to build up a huge project and have success in the end. You preferably want to create value as fast as possible and then make it more intelligent and complex implementations over time.
What makes a startup in data different?
Startups need to solve the data collection issue. They also need to solve a first use case with this data fast, otherwise, the startup will fail. Therefore, generating insight from data is usually a use case that comes before machine learning or other fancy AI solutions.
Ways to collect data
- Public data (eg. weather data).
- Data from standardized systems (eg. Satellite provider, Google Adwords).
- Generate your own dataset (eg. first, create an app that is useful to users which in return generates data from those users).
- Let users upload data with specific formats (eg. a customer file).
- Integrate with business systems (eg. get data from a data warehouse).
- Always try to minimize integration.
- Data is usually messy when it’s not from standardized systems (be prepared to spend 80% of the time working on data cleaning and integration).
- Provide a standard for your users to facilitate working with work with different customers or different businesses.
Ways of data integration
- Complete self-service.
- Standardized APIs to upload data.
- Custom integrations.