Archives September 2021

Data Science Job market in Czechia

I have done a big part of my career as a data scientist in Prague. During this time, I worked for different companies as a full-time employee and as a contractor, both as an individual contributor and as middle management. I also helped hire, and sometimes also fire, candidates. This has given me a good overview of the local data science job market that might be useful to you.

There are roughly three tiers: international companies, companies focused on the local market and startups. Here are some notes about my experience with each of them.

International companies

There are many companies that are based in Czechia but their primary market is other EU countries or the US. These companies tend to have higher compensation, from 40-60 thousand Czech crowns for entry level, and 80-100 with 2-3 years of experience. The working language is usually English and friendly with foreigners. As the main business unit or the final client is abroad, sometimes this higher pay implies meetings in different timezones and traveling. It also means that the output of the work may not be visible, and one can feel detached of it, especially in consulting companies. Since these are usually big brands, they give you a resume boost.

Local market

These are companies that have main customers in Czechia. Some of them have strong data science teams, like telecom operators (O2, T-Mobile, Vodafone), but there are many others. In these companies the pay is usually below that of international companies. They may not necessarily be English-friendly, which could be a problem for foreigners. One advantage is that the work feels (and is) closer: you can see the outputs of your models in your everyday life. They are also more involved with the local community in general, sponsoring hackathons, events, etc.


Startups overlap with the previous two, but are an important part of the local data science job market, and are a bit different. They tend to have friendlier schedules, be more open about remote work and more relaxed in general. On one hand this makes the transition from university a bit easier for many, as many startups look more like a cool student dormitory than a traditional working place. Some of them tackle riskier projects (i.e. more technically fun), and tend to have more modern engineering practices. Among these startups there are both product companies and companies focused on professional services. This second category is a smaller version of larger consultancies: they have international projects and one gets to rotate from project to project. They tend to be more open to part-time arrangements as well. There is more risk in general, if the business model is not resilient.

Is Deep Learning a valuable skill?

Many online courses and universities promote deep learning courses and bootcamps. This enthusiasm goes also to business decision makers and investors. And then, there comes the push to incorporate deep learning / artificial intelligence / machine learning models at all costs. For someone transitioning to data science, is deep learning a valuable skill to invest time and money to learn?

The truth is, unless you are working on images, audio or text, deep learning is not a very valuable skill. Deep learning excels at extracting patterns from high-dimensional data that is generated in a consistent way. Only in those cases makes sense to use deep learning. The typical data scientist that uses a mixture of SQL and classical data mining algorithms (logistic regression, decision trees, random forests) on tabular data is unlikely to benefit from it.

Even in cases where deep learning would normally work great, one should also be aware of the amount of data. If one does not have enough data, deep learning algorithms will not work. A rough estimate is 10 data points per parameter. This is a completely heuristic figure, which I cannot back up by theory. Modern neural networks have millions of parameters.

During my corporate training courses we often benchmark deep learning models against others. In most cases, deep learning models are way below the mark. One exception to this are outlier detection models. In this setting, we often get better results with autoencoders. But otherwise it is hard to make the case for deep learning.

Instead of investing time and money in learning advanced algorithms, newcomers to the field should brush up on their data gathering and analysis skills. That is definitely a differentiatior.

Should you go to a startup in your first job as data scientist?

Startups are all the rage these days, and many people are attracted by some of the benefits: lots of toys in the office, possibility of remote work, flexibility and an informal office culture. This is all fine and well but it comes to a cost. Does that mean you should go to a startup in your first job as data scientist?

Startups, by definition, do not have a tested business model. Maybe there is a need in the market, but maybe customers are not ready to pay. There will be likely a lot of iteration in the product, and that very often translates into changes in the internal processes.

This can be terrible news for someone who is starting in their career as a data scientist:

Tasks might not be clearly defined

There will be constant changes in the data models

Objectives and likely very little consistency in the day to day work.

Why is this bad for you as a beginner? Constant changes distract you from getting proficient with a single set of tools.

Ok, so should you avoid startups?

Don’t get me wrong, I think that data scientists should be proficient in many tools. After all, the work is largely the same, whether you use R, Python or a no/low-code solution like Alteryx. But the constant changes in the beginning distract you from grasping the fundamentals of real-life data science work.

Startups are fantastic places to go as your second job, once you earned enough experience in a corporate environment or a larger company. You will really appreciate the flexibility in the work, and since you will be, at that point, much more familiar with your tools, you will thrive in the wonderful chaos a startup can be. But I would highly recommend beginning your career in a more structured environment. That does not mean you cannot succeed in a startup in your first job as data scientist, but you should be aware of the risk.