Reading time: 10 min
You want to work with data, but the job titles blur together. Data scientist, data engineer, data analyst, ML engineer: in job ads they often sound alike, yet in practice they are different jobs, with different tools and different pay.
This guide explains what a data scientist is, what the job looks like day to day, how it differs from a data engineer and an analyst, which skills it takes, how to get into the field and how much data scientists earn in Poland. If you are looking for a role now, see our data jobs. ITDS also builds nearshore teams from Poland and Portugal for companies across Europe and the US, and many data science projects come from those clients.
| In this guide | |
| What is a data scientist What they do Data scientist vs data engineer vs analyst Tools and skills |
How to become a data scientist Recruitment Career paths Salaries Data science at ITDS FAQ |
What is a data scientist?
A data scientist turns data into decisions. The role combines statistics, programming and business sense to answer questions a standard report can't. Which customers will leave next month? Which transaction looks like fraud? What price keeps sales from falling?
Put simply, an analyst describes what happened, while a data scientist predicts what will happen and builds models that do it automatically. More and more often, data scientists are also responsible for getting a model into production, not just for a prototype in a notebook.
What does a data scientist do day to day?
The scope depends on the company, but a typical week includes a few recurring tasks:
- Talking to the business. Understanding the problem and turning it into a question data can answer.
- Preparing data. Finding, cleaning and joining data from many sources. This is often the biggest part of the job.
- Analysis and modelling. Statistics, machine learning, hypothesis testing and deciding which model works best.
- Deployment. Working with data engineers and MLOps teams so the model runs in a real system.
- Communicating results. Visualisations, presentations and recommendations for people who don't speak statistics.
On projects ITDS recruits for, data scientists work on things like fraud detection and customer churn in banking, advertising models in ad tech and data analysis in healthcare.
Data scientist vs data engineer vs data analyst
All three roles work with the same data, but each is responsible for something different.
| Data analyst | Data scientist | Data engineer | |
|---|---|---|---|
| Main question | What happened and why? | What will happen and what should we do? | How do we deliver data quickly and reliably? |
| Typical work | Reports, dashboards, ad hoc analysis | Predictive models, experiments, machine learning | Data pipelines, warehouses, integrations |
| Key tools | SQL, Excel, Power BI, Tableau | Python, SQL, ML libraries, statistics | SQL, Python, Spark, cloud, ETL tools |
| Works most closely with | The business | The business and data engineers | Data scientists and IT teams |
The lines are often blurred, especially in smaller companies. Many people start as data analysts and move into data science over time, and some data scientists later specialise in ML engineering or MLOps. We cover the infrastructure side in a separate guide to the data engineer role.
| An analyst describes the past. A data scientist builds a model that predicts the future. A data engineer makes sure both have something to work with. |
Tools and skills a data scientist needs
Programming and data
The foundation is Python, with its libraries for analysis and machine learning, plus solid SQL. In ITDS job ads for data scientists, SQL and Python appear almost every time, often alongside big data tools such as Spark and cloud services.
Statistics and machine learning
Without solid statistics, it is hard to tell whether a model really works. A data scientist should understand hypothesis testing, regression, classification, model validation and common traps such as overfitting or the wrong metric.
Cloud, MLOps and language models
The market increasingly values the ability to deploy models, not only to build them. Experience with the cloud, containers and MLOps tools makes a candidate stand out. The same goes for work with large language models and natural language processing.
Business skills and communication
The best model achieves nothing if nobody understands what it means. Explaining results in plain language and knowing an industry, such as finance, often decide a promotion more than one more library.
How to become a data scientist: portfolio and first job
There is no single route into the field. Some people arrive after degrees in mathematics, computer science, econometrics or physics, while others retrain from data analysis or programming. Whatever your starting point, a few steps work best:
- Master the basics. Python, SQL and statistics. Without them, no ML course will help.
- Build two or three end-to-end projects. From raw data to a working model and a short summary of results. Fewer, but done properly.
- Show your code. A public repository with a clear README tells a recruiter more than a list of courses.
- Pick a domain. Projects in finance, e-commerce or medicine show you understand real problems.
- Consider starting in analytics. A data analyst role is a common and sensible path into data science.
For mid and senior roles, commercial experience matters most. ITDS job ads for data scientists usually ask for two to four years of work on similar projects.
Data scientist recruitment: what to expect
The process differs between companies, but most follow similar stages. It pays to prepare for each one.
- Screening call. A recruiter checks your experience, expectations and preferred contract. Prepare a short summary of the two or three projects that best show your skills.
- SQL and Python test. Usually a few tasks on queries, joins and data processing. It sounds simple, yet many candidates fall here.
- Take-home task or case study. You get a dataset and a business problem. What counts is not only the result, but how you explain your choices.
- Technical interview. Questions on statistics, model choice, metrics and how to check that a model really works on new data.
- Meeting the team or client. A check on how you work with others and explain results to people outside IT.
Candidates most often focus on algorithms and skip the data and the business context. Recruiters want to see that you understand why a model is being built, not only how.
Career paths: where a data scientist can go next
Data science is a good starting point for several specialisations. After a few years, most people choose one of these directions:
- ML engineer. More engineering, deployment and scaling of models in production.
- MLOps. Automating how models are trained, monitored and updated.
- Language model specialist. Work with large language models, NLP and AI-based systems.
- Lead or principal data scientist. Leading a team, setting direction and owning the biggest projects.
- Domain expert. Deep specialisation in one industry, such as credit risk or fraud detection in finance.
Each of these paths usually raises pay. Today, rates grow fastest for people who can take a model from idea to a working system.
How much does a data scientist earn in Poland?
Data science is one of the better-paid specialisations in Polish IT. Pay depends on experience, industry, city, contract type and whether you work with production models, MLOps or large language models.
| Level | Experience | Employment contract, gross per month (approximate) |
|---|---|---|
| Junior | 0–2 years | about PLN 8,000–13,000 |
| Mid | 2–5 years | about PLN 14,000–22,000 |
| Senior | 5+ years | about PLN 22,000–34,000 |
These ranges are approximate and based on public summaries of 2026 job offers. Use them as a reference point and compare current offers before you decide.
At mid and senior level, many data scientists work on a B2B contract. According to the ITDS IT Talent Trends 2026 report, senior data science rates on B2B are usually around PLN 140–170 net per hour. When you compare B2B with an employment contract, remember the costs you cover yourself on B2B: social security (ZUS), holidays, sick leave and gaps between projects.
People who combine data science with model deployment, the cloud or experience in regulated industries such as banking gain the most.
| Looking for a data science project? We recruit data scientists for long-term projects with international clients. See data jobs |
Data scientist at ITDS: projects and recruitment
ITDS is an IT staffing and nearshore company with teams in Poland and Portugal, part of House of HR. We recruit specialists for projects with international clients, mainly in financial services, but also in technology, digital advertising and healthcare.
The process is short. You apply or send your CV, talk to a recruiter, take a technical interview with a specialist, meet the client team and receive a clear offer with the contract type and rate. Our projects usually run for a long time, and we stay in touch once you start.
Data scientist: frequently asked questions
A data scientist analyses data and builds models that predict behaviour or outcomes, such as customer churn, fraud or demand. They prepare data, build and test models, help deploy them and explain the results to the business.
A data scientist builds models and draws conclusions from data. A data engineer designs and maintains the systems that collect, process and deliver that data. The two roles work closely together.
As a rough guide, a junior earns about PLN 8,000–13,000 gross a month on an employment contract, a mid about PLN 14,000–22,000 and a senior about PLN 22,000–34,000. Seniors on B2B usually earn around PLN 140–170 net per hour. Pay depends on industry, city and specialisation.
Start with Python, SQL and statistics, build two or three complete projects with public code, and consider starting as a data analyst. Commercial experience with data makes the move into data science much easier.
Yes, at least the core models and how to evaluate them. Increasingly, the ability to deploy models to production and to work with large language models matters too.
Find your next data science projectWe recruit data scientists for long-term projects with international clients, with a clear contract from day one. See current offers and apply in a few minutes. Companies looking for specialists can learn about our nearshore teams. |