It's the same old story. A company decides it needs "Innovation" and "AI." They hire a Data Scientist (you). First day on the job: they ask you to fix a Java API or center a div in the administration dashboard.
If this sounds familiar, you're not alone.
The problem isn't bad faith, it's the terminological confusion. For many managers, "if you write code, you're a developer." And that's a very costly mistake.
Today we're going to break down what a Data Scientist actually does, why the profession is much broader than it seems, and, most importantly, What to say to your boss so that he stops using you as an expensive backend developer.
The origin of the misunderstanding: Code vs. Product
The root of the problem is simple: both the Data Scientist and the Backend Developer use similar tools (Python, SQL, terminals, Git).
But the aim It is radically the opposite:
- The Software Developer (Backend): He builds systems. His world is deterministic. If the code is well written, output A will always produce result B. Its goal is system stability, scalability, and efficiency.
- The Data Scientist: He seeks answers. His world is probabilistic. Experiment with data to reduce uncertainty. Your goal is not (always) to build a final software product, but to extract value, patterns, or predictions from the information.
If you put a Data Scientist to do pure Backend tasks, you're using a microscope to hammer nails. It works (sometimes), but it's a waste of resources and you'll end up breaking the microscope.
The range of the Data Science (We are not unicorns)
Another point to explain to your client or manager: "Data Science" is an umbrella term. Within this field, there are very clear specializations. Expecting a single person to do everything is a recipe for burnout.
To make it easier to understand, divide it into three profiles:
- Data Analyst: Look to the past. Answer "What happened?". Use SQL and visualization (Tableau/Power BI) to explain the business.
- Data Engineer: Build the pipelines. Ensure that the data arrives clean and fast. This is the closest thing to traditional software development.
- Machine Learning Scientist: Look to the future. Answer "What will happen?". Create predictive models and algorithms.
When a company hires a generic "Data Scientist" and asks them to build the infrastructure (Engineering), create dashboards (Analytics), and train models (Science), they usually end up doing none of that and just patching legacy code.
The script: How to explain to your boss what you do
Here are the key points for your next one-on-one meeting. It's not about complaining, it's about speak their language: profitability and efficiency.
1. "I sell uncertainty reduction, not lines of code"«
Explain that your work is scientific. You have a hypothesis (e.g., «"We can predict which customers are going to leave."»), you experiment and get a result.
A developer receives clear specifications to build something that we already know that we want. You discover what we don't know.
2. "My life cycle is different"«
Software development is usually linear or agile with clear deliverables. Data science is iterative and exploratory. Sometimes, the result of a week's work is discovering that an idea is not viable.
And that's valuable: you've saved the company months of development on useless functionality. That's what you do.
3. "If I do backend development, you'll lose money"«
Be pragmatic. A Senior Backend Developer is usually faster and more efficient at writing APIs and microservices than a Data Scientist.
If you dedicate yourself to that, the company is paying a specialized salary to obtain a mediocre result in an area that is not yours, while the data remains there, untapped, losing value every day.
Conclusion: Specialization is maturity
The Data Scientist career is fascinating because it blends statistics, programming, and business acumen. But for it to truly shine, it needs an environment that understands that We are not "the guys who fix the computer" or "extra programmers".
We are the ones who take the company's most valuable asset (data) and turn it into smart decisions.
If your boss asks you to change the color of a button "because you know how to program," send them this article.