5 Lessons I Learned My First Year as a Data Scientist in a Startup

data scientist lessons learned Oct 29, 2022
Sarah wearing a black dress with dots in a field full of flowers

Working at a startup is not for everyone. Ideas are pitched, teams are put together, and products are launched at the drop of a hat.

I am so glad I had the opportunity to be a part of a couple of startups as either a data scientist or a data engineer. It has changed my life, and I have learned so much. I have learned about machine learning, data science, and statistics as I have gone from working on a single project to multiple projects in different companies.

I have seen my growth in terms of what knowledge I have gained, and I want to share these learnings with you today.

1. You have to become independent. Quickly.
Startups are small, and most companies will not focus on data team hires. Hiring the first person on a data team is a scary thing, but not as frightening as realizing you are literally on your own.

I became a data scientist in that situation. I had a little over a year of experience as a data engineer under my belt. My friends will tell you I do not shy away from a challenge, so when this opportunity came my way, I had to grab it.

My first significant roadblock was figuring out how to manage projects in addition to my load. I used to have project managers to help me, but now I have to learn how to prioritize my tasks, set up requirements, and create proposals for new data pipelines or tools.

My second major roadblock was sharing the details of what I worked on with not just my manager but also the stakeholders. I presented with CEOs and upper-level management and created a story in emails with the statistics I had recently calculated for the whole company.

But through those roadblocks, I grew as an individual. I became more independent and learned how to solve problems quickly.

2. You will do all the data jobs.

I worked with a great data analyst. He was smart, knew his way around Tableau and reports, and could handle anything you threw at him. But it was the data engineering that got to him. How do we get the data from point A to point B?

The company might have hired you as a data analyst, but specific projects will require data science or engineering. As a data analyst, you will learn how to do those tasks. And the same goes for data scientists and engineers.

In startups, data is extensive, and there is no specific discipline or expertise which a data person needs to know. More importantly, every data person should be knowledgeable in different industries because this will make them an invaluable resource for the team.

3. You will quickly learn what you like and do not like.

This one goes hand in hand with number 2. Because you will become every position in the data world, you will gain an unbelievable amount of experience in tools you did not want to learn.

I love learning new things, but sometimes I wish I could avoid learning. Occasionally, clients request something specific that you do not want to know. For example, a client might ask us to create a dashboard in PowerBI, so I learned it and completed the task successfully. However, I did not have any specific plans for learning PowerBI — it was for that one project.

Those experiences are invaluable, though. You will move on from your current position and find positions without that specific tool.

4. You will need to set up the processes on your own.

A data team must set up data and code standards to ensure best practices. These processes include code reviews, creating CICD pipelines, and setting up team meetings.

Unfortunately, most of these processes are not in place or are in place to meet the needs of other teams. You will need to figure out what is important to you and implement the most critical processes within reason.

When I worked at Qumulo, I got hired to build up our processes as a data team. Out of all the missing processes, moving our data pipelines from on-premise to AWS provided us with the most benefit. And thus, I started by setting up Gitlab and later infused CICD pipelines to get our code in AWS Fargate instead of manually moving our code to our production environment.

5. Communication

You will have the opportunity to work closely with many teammates who do not have extensive knowledge about data. It is your job to devise solutions for these tasks using your initiative. You will also learn how to speak about these solutions in layman’s terms so that people who do not have a background in data can understand them.

That skill — translating your work to non-technical folks — is vital. I have had a product manager, a vice president of Product, or even a Chief Product Operations as my manager. My managers meant well and tried to understand what you are saying but know enough about tech-related terms to be dangerous. Undoing those definitions and further explaining to them what is needed requires skillful communication.

Conclusion

Being the first data scientist at a startup has given me new opportunities. I learned how to act quickly and was exposed to other parts of the business. I also found out that PowerBI is not something I want to focus on more. Meanwhile, I am still learning how to build processes and teach a non-technical audience what and why I am exploring.

As always, keep teaching and keep learning. I will see you next time!

Stay connected with news and updates!

Join my mailing list to receive the latest news and updates.