ML/Data Career Guide for Newbies

All you need to know to successfully launch your career in data: today's essential professions in the fields of data science, machine learning and AI. By which criteria employers select candidates for entry-level data jobs. And how you can discern the truly interesting job openings.

This guide is based: (i) our personal experiences of studying data science and entering the job market (our own and our peers'), (ii) the insights we gained into the organization of different companies from working as data consultants, and (iii) on an analysis of current data job vacancies in Germany.

TLDR

Data career optimization from the get-go (in 5 steps)

1. Educate yourself. About the different data and ML/AI professions and what's working in these professions would be like (which tools? what kind of projects?).
2. Specialize. If you are about to enter university or you're in middle of your studies, choose courses that match your career aspirations.
3. Educate yourself even more. About potential employers (do they have a CTO? How serious are they about digitalization? Would you be their first data hire?)
4. Fix up your CV & and GitHub profile. Log your university projects, Kaggle competitions etc. Go intern. Work as a working student ("Werkstudent").
5. Apply for jobs. Even if you don't meet 100% of the requirements (demand for data experts > supply).

The top-6 data jobs in Germany

Data scientists, data analysts, data engineers, ML engineers, ML scientists, solution architects, and data analysts are the most sought-after data experts on the German job market.

Data scientist

• What does a data scientist do? mostly data analysis (mostly with Python/R).
• Hotness: in extremely high demand (keep in mind: some data scientist vacancies entail more data engineering than data science responsibilities. The correlation between job title and actual tasks isn't perfect.)
• Related professions: analytics/BI engineer

Data engineer

• What does a data engineer do? set up and maintain data infrastructure. SQL ist their lingua franca.
• Hotness: in high demand (data engineers with cloud expertise even more so)
• Related professions: cloud engineer

Machine learning engineer /AI engineer

• What does an AI/ML Engineer scientist do? develop solutions and products based on AI/ML/LLMs.
• Hotness: in very high deman. Especially in the US, ML Engineers are very much trending right now.
• Related professions: AI/ML developer

Machine learning scientist

• What does a machine learning scientist do? machine learning an AI research.
• Hotness: hottt. OpenAI, DeepMind, Meta, etc. are all hiring ML Scientists.
• Related professions: ML researcher.

Solution architect

• What does a solution architect do? develop and implement tailored end-to-end solutions. Solution architects know their way around data and software.
• Hotness: steaming hot. They are versatile and render things possible, that are impossible to achive with cookie-cutter tools.

Data analyst

• What does a data analyst do? data analysis in Excel (sometimes confused with data scientists. But data analysts rely on an entirely different toolkit).
• Hotness: solid option (also suitable for lateral recruits).

These are the requirements for a data career in Germany:

The majority of data jobs in Germany require a university degree, i.e. at least an undergraduate degree (BSc). . For research-heavy ML/AI jobs, employers expect postgraduate degrees (MSc, PhD).

The best study programs to prepare for a career in data/ML

The majors offering the best preparation for a career in data and machine learning are: computer science, data science, maths/statistics and related study programs.

Less obvious, but apt choices include: economics, sociology, or psychology (if these programs emphasize quantitative methods).

Minimum tech skills for ML/AI and data jobs

In addition to getting a suitable degree, you should also be familiar with programs, programming languages and packages most commonly used in the respective data job:

• Data scientist: Python (Pandas, Jupiter Notebook), Excel, SQL, Apache Superset/Tableau/Power BI, ML
• Data analyst: Excel, Python
• Data engineer: SQL, Python, databanks, Cloud (AWS/Azure/GCP)
• Machine learning scientist: PhD-level ML research, PyTorch
• Machine learning engineer: ML, PyTorch + data engineer skills
• Solution architect: a bit of everything (ideal combination will be domain-dependent).

How do I find out which data career suits me?

Data and ML professions differ not only in their tools and skills, but also in their mindset and approach to work. Already as a student, it's worth exploring these differences. There are (at least) 4 ways to get a feel for the specificities of working as a data scientist, ML engineer and the likes:

1. Attend lectures and seminars that expose you to a broad range of topics. This is a good call, especially when your study program covers various data sub-disciplines: data science programs, for example, have data engineering and machine learning classes in their curriculum. (If you've already figured out your dream job, say ML engineer, then specializing hardcore–in this case on machine learning–would also work great.)
2. Improve your coding game in Kaggle Competitions (Kaggle = online plattform for data science competitions/hackathons).
3. Do internships or become a working student ("Werkstudent").
4. Learn from the experience of others on twitter/X, blogs, in newsletters and podcast. Ask your class mates about their internships, working student gigs, etc. Or join an ML/AI meetup (most larger cities have one, there are online meetups as well).

Where to apply to? First Data Job Heuristics

With a background in data science, machine learning and/or artificial intelligence you can choose between two basic types of jobs: jobs focused on applying technologies and jobs aimed at research. But that's only the most basic distinction. There are more..

Not all data jobs are created equal: 5 types of companies to be aware of

More and more companies want to hire data experts. The choice is yours. But how to determine which company would be a good fit for your preferences and career goals?

Before starting as a junior data professsional at some company, you should find out whether data endaevors and digitalization are considered a prioroity there. The company's data philosophy will affect what kind of projects you work on, the tools to your disposal, the team you'll work with and your opportunites to pick up new skills.

Based on their respective data savviness, we can distinguish 5 types of employers/companies:

1. Tech companies
   • Data philosophy: software/data = core business; there's a CTO (software/data).
   • Good fit for: young data professionals who want to work on highly specialized tasks and who appreciate promotion prospects all the way to the top (C-level).
   • Examples: SAP, Zalando, N26.
2. Non-tech company w/ data department(s)
   • Data philosophy: software is not their core business, but they prioritize digitalization.
   • Good fit for: junior data experts with a preference for well-defined responsabilities and career levels.
   • Examples: BMW, BASF, adidas.
3. Company w/o data department
   • Data philosophy: you = first data hire.
   • Good fit for: DIY-spirited young professionals with robust programming skills under their belt. As first data hire, you'll have to set up systems from the ground up. This means working independently on an array of (often changing) tasks.
   • Examples: small and medium-sized enterprises that, up until now, have outsourced all their data-related work.
4. IT/data consultancies
   • Data philosophy: data consultancies implement software solutions for business clients of various sizes and industries.
   • Good fit for: young professionals who like working on varied projects and learning about the data needs of different companies.
   • Examples: Accenture, Capgemini, Deloitte.
5. Tech-Startups
   • Data philosophy: software/data = core business.
   • Good fit for: creative young professionals who appreciate a fluid work environment and challanging, frequently changing tasks.
   • Examples: Black Forest Labs, Osapiens, TradeRepublic.

In case you cannot automatically tell a company's type, check their webpage or their LinkedIn. Job postings by the company can provide insights as well: sometimes they'll explicitly mention the professional background of the colleques you'll be working with. And pay attention to the information regarding the tech stack: modern programming languages and tools are a good indicator for a promising data philosophy, interesting responsibilities and opportunites to grow as a data specialist.

Don't hesitate to bring up a company's data strategy in job interviews. This information will help you decide whether you want to work there and you indicate interest in the company,

Research jobs in data, machine learning and AI

Many german universities have postdoc and tenured positions for machine learning/AI and data science. Private research labs offer further academic career opportunites, e.g. at Google DeepMind, OpenAI, Meta FAIR, or Stability AI. Such research careers usually require a corresponding PhD degree.

What's the entry-level salary for data, machine learning & AI jobs in Germany?

The median salary of young data and ML professionals in Germany is 62.000 € (gross).

Junior data engineers make between 45.000 € und 87.000 €. The entry-level salary of data scientists varies more: 40.000-188.00 €. Junior machine learning engineers make 67.000-100.000 € and solution architects 66.000-150.000 €.

How to apply? Tips for newbies

No professional experience to speak of, but... Newbies to the job market have to get creative, to set themselves apart from other applicants. To employers, your degree is the most important indicator that you're competent in your field. In addition, they're usually interested in your hands-on (coding) experience. When you apply for a research position, your PhD specialization and your research up until now will matter a great deal.

Aside from your diploma, you can leverage a portfolio of past projects to stand out from the crowd:

7 Strategies to win over employers as a rookie

1. Kaggle: top rankings in Kaggle competitions are a surefire way to convince employers of your data science capabilities. Those holding the title of Kaggle grandmaster are all but certain to get any data scientist job they apply for.
2. GitHub (underrated)!: the obvious choice for showcasing your project portfolio: university projects, code from Kaggle competitions, etc.
3. Certificates: spend your time (and money) only on reputable certificates, such as certificates for cloud services (AWS, Azure, GCP) or data visualisation services (PowerBI, Tableau).
4. Internships: to gain hands-on experience in general. Or to check out a specific company you're considering.
5. Working student: see internships.
6. Student assistant to a professor: to get an idea of what it's like to work in academia. Did it. Loved it (usually easy money).
7. Hackathons: especially if you plan on applying to the company hosting the hackathon.

What's the application process for entry-level data jobs like?

Every company will want to see your diploma and your CV. What happens next, can vary: perhaps a technical assesment/white board interview as a form of prescreening (especially with big tech companies), next one or multiple rounds of traditional job interviews.

4x Career advice: how to leverage your first job for your future career

1. Job titles aren't everything...
   Sometimes data job titles are being mixed-up and used interchangeably. Some job posts for data scientists come with prerequisites that a data engineer is more likely to meet than a data scientist. So always mind the description of responsibilities and the tech stack in job posting. You'll have abetter idea of what to expect.
2. **... but they stills matter.**responsibilities
   The data industry is continuously generating new trends. That's the case for job titles as well. Not all employers are able to keep up. Some use "data scientist" und "data analyst," or "cloud engineer" and "system administrator" synonymously. Even if it's the same kind of job, they will let you appear in a different light, respectively: data scientists and cloud engineers give the impression that they work with SOTA technologies, while data analyst and system administrator can appear a little dated. In short: up-to-date job titles will look better on your CV.
3. Go for it.
   Demand for data experts exceeds supply. Even if you don't meet all the requirements for a particular job, consider applying nevertheless. You might just get lucky! Especially if you can check all the "must-haves" and are only lacking in the "nice-to-haves".
4. Your first job won't be your last.
   Aside from a good salary and interesting responsibilities, many rookies also care about the career paths a job might open up.
   In general, the larger the overlap between two data professions, the easier it is to switch between them. For example: a data scientist who knows SQL could fairly easily switch lanes and become a data engineer. As to switching companies, leaving a big tech company to join a medium-sized one (perhaps in an executive position) will be easier than the other way around.