Karina Kato
💻 Data Scientist Specialist at iFood
Hello, I'm Karina Kato, from the iFood Data team. I wrote this article thinking about sharing some learnings and challenges. I have given a lot of mentoring during my career, both to people who were just starting out and to those in career transition. I know we are living in a time of information overload and it is easy to feel lost or overwhelmed in a new area like Data Science. Therefore, I decided to start a series of articles to try to help those who are going through these situations.
Before starting I would like to make an analogy with the book: “Alice’s Adventures in Wonderland” by Lewis Carroll. There is an excerpt where Alice talks to the cat asking what would be the best way forward. He says it depends on where she wants to go. When she responds that it doesn't matter much, the cat suggests that it doesn't matter which way to go then. I really like this part, because it’s like that with mentoring too. It's very difficult for someone to be able to help you without context. If you don't know your goals yourself, the tips given will be superficial.
My purpose with this series of articles is to teach you how to define your short-term professional goal and how to plan to achieve it. I also want to help you understand what it means to be a Data Scientist. You may even discover sooner that you are better suited to another data profession. It is worth mentioning that you will not be ready to be a data scientist just by reading this information. You'll have to study hard and work hard to make this happen, but I hope this series helps you do this in a more objective way.
Well, let's get started! I created the following flowchart to try to “give you a guide” so that you can not only define your goals, but also create a plan to achieve them.
1 — Understand the different data professions to choose the one that best suits your profile:
One point I wanted to highlight is that there may be differences from one company to another. Often, in companies with a small data team, professionals may have to take on responsibilities from another role. For example, it is common in some companies for the data scientist to have to perform part of the responsibilities of a machine learning engineer. So, here I will provide an overview and I recommend that after you choose which of these professions best fits your profile, try to find out what this professional does in companies where you want to try the selection process.
To also make it easier to understand the differences between data scientist, data analyst and BI analyst, I created this representation of the types of data analysis. There are 4 types: descriptive, diagnostic, predictive and prescriptive. You will see that these professionals work to generate knowledge through data that will help in decision making, but their focuses are different.
- BI Analyst (Business Intelligence Analyst)
Responsible for transforming data into knowledge that will help make business decisions. The BI Analyst works more with descriptive analysis. It has a strong focus on structuring reports, from data flow to visualization. Works with structured data.
- Data Analyst
It is a newer profession than BI analyst. Both are responsible for bringing knowledge from data to assist business decision-making and work with structured data. However, the data analyst generates descriptive or diagnostic analyzes and looks for patterns to answer business questions that can then be turned into reports for BI.
- Data Scientist
The data scientist also works with data to assist in business decision-making, but with a focus on better understanding the future. It performs predictive or prescriptive analyzes (for example, creating machine learning models) that can even be incorporated into the company's product. Furthermore, the data scientist can work with unstructured data (data that is not so well defined). He works closely with the machine learning engineer. It is also close to the business team, but not as close as a data analyst and a BI analyst. Finally, the data scientist generally has more skills focused on software development than the other two professionals.
- Machine Learning Engineer
This professional is the one who implements MLOps (Machine Learning Operations), speeding up and facilitating the process of putting machine learning models into production by creating tools and pipelines. These machine learning pipelines help maintain the health of models created by data scientists so that they have continuous monitoring, updates, and avoid deterioration in model quality. Machine learning engineers have skills similar to software developers, but with a focus on developing machine learning tools. They often work closely with data scientists and data engineers.
- Data Engineer
Responsible for the quality of the company's data. This is who creates the data pipeline (planning, development, modeling and monitoring). It is the pillar for starting a data team, as it is the professional who takes care of the data infrastructure that will be used by other professionals. Often works closely with Software Engineers, Machine Learning Engineers, and Data Scientists.
2 — Remember that there are several career levels for each of these professions:
- Trainee: student at the beginning of his career. Employment contract with different working hours. Greater focus on learning than delivery. You receive a lot of support to develop your activities.
- Junior: is someone who has recently graduated or does not yet have much experience in the field. Generally, they receive activities with low complexity and have support from other people.
- Full: has good experience in the area. You can work on more complex activities, but you still need projects to be more structured.
- Senior: has a lot of experience in the area. Can structure and develop complex projects. He can also support well and be a reference for the team.
After senior there may be other levels depending on the company's career format. For example, in companies that operate with a Y career path, there is a division into management and technician. I chose a technical career. My level today is expert. There may be differences in the name adopted by each company and level of granularity within the same position, but this is another matter.
3. Discover the skills needed for the chosen career and level:
Try to study what is necessary for your level and career. Start looking for vacancies in companies that interest you, look at what the requirements are for your level and, at most, what are the differentiators for you to stand out. Don't try to embrace the world all at once. Have focus! Research before you start planning. If possible, also talk to people who work in the area.
Study and Selection Process
After completing the previous steps, create your study plan with goals and deadlines. When you feel prepared enough, start making selection processes. Do, at most, two at the same time so as not to burn through your vacancies at once and be able to dedicate yourself to doing the process well and getting to know the company you are applying to.
Finally, even if you have already failed a selection process, don't be discouraged. Ask the company for feedback. Many technology companies are very open to giving feedback on points of improvement they noticed during the process. Add these points to your study plan too. If it is still difficult to enter the market, look for companies and vacancies with lower levels of demand until you gain more experience. Sometimes, we also have to know how to take a step back to take two steps forward later. Success is not a straight line.
I hope this introduction helped give you some direction, see you in the next one!😊
Follow me to receive notifications of upcoming articles. What's next:
- Valuable tips for aspiring data scientists — introduction ✔️
- Valuable tips for aspiring data scientists — concepts
- Valuable tips for aspiring data scientists — skills
- Valuable tips for aspiring data scientists — study materials
Feel free to leave your feedback here and share this article with anyone trying to break into the field.
*I would also like to thank Gabriel Campos, Jhones Pinto and Jaime Kuei for their tips and feedback on this article.
