The world’s leading publication for data science, AI, and ML professionals.

5 Technical Behaviors I’ve Learned from 2 Years of Data Science and Engineering

Sometimes you need to take a leap of faith and switch careers

Photo from Jens Lelie on Unsplash
Photo from Jens Lelie on Unsplash

It is crazy to think I am coming up on two years of being in Data science and engineering after my sudden career shift away from DevOps. In 2017, I began my Masters in Data Science. I took a leap and left a stable job to go back to school to study what I believed was to be my passion. I have worked as a data engineer and a data scientist within two different companies in the past two years. I have learned skills in data ingestion, cleaning, analysis, and more. As I approach the 2-year mark, I want to sit back and do a retrospective of what I have taken away from this journey so far, with one burning question in mind:

Did I make the right move into data science?


As I began my journey into Data Science, I have worked on both large and small projects. After about one year worth of work, I realized the value of feedback in the workplace.

1. Get Feedback

Getting feedback on the direction you are taking and the conclusions you are making from the data can help you decide on your next steps. This feedback can vary from having a colleague look over your code in a code review to having a subject matter expert examine your output to determine if there are glaring issues with the analysis. No matter what type of feedback you are looking for, learn from the comments that are being made and determine how to use that moving forward.

Tip: Be open to receiving feedback. You don’t need to accept all the feedback you get, but listen to what your colleagues have to say and determine what you will take from that conversation to work on.

To continue getting feedback for my work, I will host 1:1 sessions or reach out to colleagues to review what I am doing. Having these discussions, especially after completing large projects, has helped me understand what I am doing well and where I can improve.


The large projects are the ones I have learned the most from this past year, though. These projects span multiple months and, many times, numerous team members. My biggest takeaway from starting any large project is to develop a minimal viable product (MVP). I have found this helpful with software and data science projects as you can get an initial direction for the project and start getting feedback.

2. Develop an MVP

Whether you are creating a dashboard or developing a software library, creating an MVP first and getting feedback will help make sure you are going in the right direction with the product. It is easier to ditch something that isn’t working if you didn’t spend too much time developing it before getting feedback.

Tip: Learn to iterate quickly and fail fast on projects. If the work begins to go in an unwanted direction, stop it before too much has been invested.

Through these MVP’s development, I can decide if this is the correct direction that my work should be taking or do I need to take a step back and pivot. On my team, I also am in charge of developing the Tools and Machine Learning Platforms roadmap. Placing in MVP’s and brief research projects has helped determine if our roadmap needs to shift directions to keep aligned with what is or is not working.


My biggest accomplishment would be my team’s documentation through the development of these projects and research. I am not an engineer or data scientist who takes documentation lightly. I believe that documentation should encompass as much as possible, from API and code documentation to architecture diagrams and how-tos.

One of my first projects at my current job was to develop documentation for the team. This documentation allows someone to onboard quickly into the team and get started on projects as soon as possible. If you are beginning to develop your analyses and projects, I recommend documenting your next person’s work.

3. Write Documentation

If you are developing software or new processes, write detailed documentation that explains the architecture, the changes, and how to implement the work. This documentation will be beneficial as you bring on more data scientists or engineers into your team. It is a starting point when onboarding and can help clear up much confusion.

Tip: Learn to write documentation for your projects. It will help you and others pick up where the work has left off, especially if the work is not continued for several months or years.


Learning to document and read code are two valuable skills I took for granted during my college years. Taking a step back, I wish I focused more on properly documenting my code and reading others.

4. Read Code

As much as you are writing documentation and presenting tutorials, it would be best to read code. Whether it be from your own team’s codebase and analyses or others on the web or Github that you find interesting, reading code can be a valuable skill to have. It will aid you if you take over someone else’s work, take on tasks for a new project in a different language or codebase that you may not be familiar with, or are presented with a new API or library you need to use. Being able to understand what you are seeing will help as you decide what to do next.

Tip: Reading code may sound boring, but it can be beneficial to you. If you don’t know what to read, consider reading through a colleague’s code or some library functions that you often use.

Reading code will give you more insight into your coding style and how others code. This past year, I have spent so much time reading others’ code from within my team that I can identify who has coded a specific portion of the codebase based on stylistic choices. It is fascinating to see how you can write the same functionality in different ways.


Lastly, if these past two years have taught me nothing else, they taught me to take a step back from coding and walk away.

5. Breath and Take a Step Back

When working on an analysis or developing some software or tool, it can be common to keep chugging away on the project. You can often find yourself hours deep into something with no break, no stop, and sometimes frustrated with a bug or an issue. Take a step back from the project you are working on, even if it is only a five to ten-minute break. Walk around the house, grab the mail, or even fill your water bottle up. These breaks are essential and sometimes can help your project more than cramming away hour after hour. I often find my best ideas or solutions to come from taking a step back from my code and returning after having cleared my head.

Tip: If you find it hard to fit in a five to ten-minute break on the calendar, then schedule yourself one. Put meeting blocks on the calendar that allow you to step away for a time. You can even do this for lunch or large blocks of time to have focused development time.

Allowing myself to have these brakes in between my coding sessions has helped spark new ideas, solutions to problems, or get away long enough to allow myself to refocus. When I get back into coding, I am at my best to solve my problems and fix bugs.


Final Thoughts

This article dives into how my technical skills have developed these past two years since transitioning into data science from DevOps. These past two years have taught me a lot in terms of technical skills valuable to have in data science. I have continued coding, but with projects focused on tools and machine learning applications within the data science field. To recap, I have:

  • Developed better habits and techniques to gather regular feedback.
  • Learned to cherish and ask for an MVP before diving deep into a project.
  • Grown fond of team documentation for code, architectures, and how-to’s.
  • Effectively and efficiently read code and understand stylistic differences.
  • Take a step back from the code to take a break and regroup.

So the real question is,

Did I make the right move into data science?

In terms of technical skills and continuing to develop myself in both software and data science, the answer is yes. Transitioning into the data science field has allowed me to develop these skills and more.

In the next part of this blog series, we will check how this transition has also affected my leadership skills. And we will discuss if moving to data science was the right choice in terms of leadership development or not.

Do you feel you have the ability to develop technical skills within your data science or engineering role? Why or why not?


If you would like to read more, check out some of my other articles below!

Communication Can Make or Break a Data Science Project

3 Programming Books Every Data Scientist Should Read

15 Topics to Consider as You Review Code in Data Science


Related Articles