Machine Learning Zoomcamp: Free ML Engineering course. Register here!

DataTalks.Club

Starting a Career in Data Science at 45

From microbiologist to data scientist

20 May 2022 by Tatyjana Ankudo

A citizen of Belarus Tatiana Ankudo left the country 20 years ago. A graduate of the Biological Faculty of the Belarusian State University, she worked as a microbiologist in laboratories in Minsk, Hungary, Sweden and Germany, received a PhD in biology, Hungarian citizenship, got married, gave birth to a daughter. And a few years ago she changed her life and became a data scientist. Tatiana has been working for the international company dunnhumby for 7 months, analyzing data from a large grocery store chain and growing as a data analyst. We talked about what it is like to go into the IT field when you are over 40 and start building your career over again.

“The first two months I cried and said - its not for me”

Why did you decide to get a new specialty, where you have to achieve everything from scratch?

I couldn’t continue growing as a microbiologist. Reagents, equipment - all this costs a lot of money, and Hungary is not the richest country. It seemed logical to choose a new path: at the last job I generated a lot of data, and I had to to process it. Suddenly it turned out to be more interesting than experimenting.

It’s one thing to try, another is to change your profession. Was it scary for you?

I always tried to live so that fears do not influence my choice. The thought of leaving was frightening - I wanted to be a biologist since I was little. I took half of a year to make a decision about my profession change.

Trying to enter the IT sphere, young people in Belarus, where you are from, take three-month courses and try to get a job. It’s enough?

To go to DS - no, you need knowledge in several areas - programming, domain area, mathematics. They study programming only for at least six months.

How much did you study, and what was your path?

The first step was to enroll in Java courses, which were held offline. For 5-6 hours a day, 5 times a week, we studied with teachers. It cost about 500 thousand forints a month - this is about 1.5 thousand dollars. Then, for about the same amount, I bought a two-year online DS course.

I always liked computers and programming, and the courses made it clear thatI could do it on a professional level. The first two months I cried and said - “it’s not for me”, but then it started working out for me .

Why did you cry?

Programming requires a certain mindset. When you learn your first language, there are many rules and conventions to follow. Immersion in them is incomprehensible. It’s very hard to remember it all. You understand what the teacher is saying. But then a teacher gives you a task and you have no idea how to do it. You have to think, remember and think quickly.

What were the main difficulties in mastering DS?

To understand the tasks that are set for you. Requests come from real customers who form the question, adapting to themselves. They say one thing, but mean another. And you must understand them. Sometimes you can ask what exactly is needed, but people often get annoyed with clarifications.

Isn’t it too much to enter the specialty - 2 years?

The courses allow you to have a job and study in parallel. 6 hours a week is enough to work through the materials at a very good pace.

The knowledge acquisition process in DS should be viewed as circular. You go through linear algebra, statistics, programming, and after a while you return to them, deepening your knowledge with each cycle.

My employer is paying for Udemy courses and giving 2 days a month for self-education, training is prescribed in my career plan. DS is also about communication with a large group of people, meetings, conferences, competitions and constant growth.

In the course I took, there was a career module. After a year of study, you can start looking for a job. I went through this course in 5 months and started mailing out resumes. I wrote to about 60 companies, giving preference to those who worked in biology. They were interested in talking to me, but after the test tasks, they rejected my application

“My resume had work experience”

How many interviews did you go through?

About 10. This is a good response rate - I had work experience, which was given by a six-month internship in an American company: during my studies I was offered to work with clients on real data for free.

Did it take a lot of time to find a job?

2 months and 4 interviews. First, I completed 2 test tasks on knowledge of programming and theory, then there was a meeting with the employer - all in an online format. I was tasked with preparing a report, presentation and storytelling for a fictional company with fictional problems.

It was strange to talk about a non-existent company, but at the beginning of the interview, the employer began to actively nod, showing that he liked my train of thought and I was going in the right direction, half an hour of slideshow turned into a pleasure.

The fourth interview was a kind of face control: an employee of the company found out if I fit their team as a person. I went over.

“People aged 40+ have advantages over young people”

What from your previous experience helped you change your career?

Life experience. When, in addition to studying at the university, you have 3 years of graduate school, 2 years of postdoc and 10 years of work experience, you develop self-confidence and a professional approach to life.. After getting a job, I found out that they selected me among 200 other applicants. The employer took into account my work experience, scientific degree and soft skill I gained through the long career path.

A number of Belarusian companies are looking for young people - it is easier for them to enter the specialty.

I was looking for a job in Hungary, where ageism issues fade into the background. On the contrary, people aged 40+ have advantages over younger ones. The age problem is typical for Russia, when the age in the resume becomes an obstacle. But this is not a typical problem in Europe.

Without what knowledge it is impossible to start in DS?

Programming. Even though you can do a lot of data processing in Excel, not knowing how to program severely limits your options. Storytelling is also important: you need to communicate the results of your work in an accessible language. If a great programmer cannot explain what they did, in a simple language, the value of their analysis is low.

What level is expected of an applicant for a DS job?

Sufficient knowledge of programming and Python, and some knowledge of mathematics - at the school level. You also need to be able to communicate with others - it is quite difficult for introverts in analytics. There is a fundamental part of DS, though, when there is a requirement for communication.

What skill was the most difficult to master?

Everyone is afraid of mathematics. I had this barrier as well: when I saw a formula, I didn’t know what to do with it. My fellow students helped with it - a guy from Russia and a girl from Australia. I noticed that the material in the 2-year courses was not good enough, and I invited them to study additional material together on our own. I came up with a lesson plan and suggested a format: listen to the topic, make a presentation about once a week and answer questions.

We were a team, pulling each other, and no one wanted this to end. When someone was not ready, the next week they had to give 2 lectures. We completed the course in six months and were very proud of it.

Have you met people who tried to switch to DS, but failed?

Yes. Some students did not like the quality of the material, they took the money back, continued their studies independently and got lost in the wilds of free courses.- They didn’t have enough self-organization. Others switched to full-time courses, and this added another 2 years to their studies. Others just disappeared..

“To be specialists in three areas”

In what areas is the analyst most in demand?

Wherever there is data. The most developed analytics is in the financial sector, work is underway in the product segment. In medicine and biology, everything is just beginning.

What direction do you think is the most promising?

Pharmacology and the search for new drugs. For example, a task arises: find a cure for a cold. Scientists need to carry out a thousand experiments and a lot of time. The analyst simulates part of the results on a computer and offers 50 experiments, this is already a solvable problem.

Should such an analyst understand pharmacology?

He must be a specialist in three areas. When they explain what DS is, they draw three overlapping circles - knowledge of the subject area, mathematics and programming.

Are you planning to use your knowledge in microbiology?

Certainly. The result can be successful work in a pharmaceutical company: I understand the business problem, read scientific literature, communicate with scientists in their language.

Do you need to control the work of the analyst?

Yes, because the success of the company and financial investments depend on his conclusions. Therefore, we analysts double-check each other’s results. To be able to admit a mistake, to be glad that it was found on time, and to go forward is an important quality in DS.

Magic or analytics

DS specialists receive the highest salaries in the IT field. Did you come here for money?

Rather, for opportunities. The financial sphere attracts, but the tasks are also interesting. People always wanted to control their lives, went to fortune tellers, laid out cards. DS tools allow you to predict the course of events on your own - and no magic.

What advice do you have for a newbie who has decided to work with data analytics?

Don’t give up and go forward. I often hear questions: I’m already 30, isn’t it too late to change my profession? I am 45, and I made it in time. In the first half of my life I made a career as a microbiologist, in the second I will achieve success in DS.

Subscribe to our weekly newsletter and join our Slack.
We'll keep you informed about our events, articles, courses, and everything else happening in the Club.


DataTalks.Club. Hosted on GitHub Pages. We use cookies.