How to Think as a Business Data Scientist

4 min readMar 6, 2023

Our guest today is Ashish Patel, Chief Data Scientist at IBM. He’ll share his story, how to get into the field and best practices to follow as a business data scientist. He’ll also tell us why the agile methodology is not the best to use for ML/AI projects, what are the important things to follow to make your data science project successful and why failure is welcome.

The data science team needs to understand thoroughly the key decisions that the business users are trying to make. Then the team can present where and how the analytic results can help the business users make better decisions. To ensure that the analytic results are relevant and meaningful to the business, it is also critical to tie them back to the organization’s key financial or business drivers.

This and more is what Ashish will teach us in order to think as a business data scientist. Please stay tuned.

You will want to hear this episode if you are interested in:

[00:01] Introduction to the guest: Ashish Patel
[01:07] Ashish’s hobbies and fun facts
[02:35] Ashish’s journey to becoming a chief data scientist at IBM
[05:58] Suggestions and advice to anyone who would like to become a data scientist on how to approach everything
[11:26] How often do you have to upgrade to new technology to solve problems
[12:53] Instances when to give up on being a researcher and move to something else
[15:33] Management and the business side not understanding the research
[17:38] How to buy more time from a client in case of a machine learning solution is not up to the mark or it’s not working
[19:17] How to align business objectives to an evolving AI/ML project
[22:09] Why agile methodology doesn’t work well for data science projects
[24:40] Understanding the business
[25:54] What’s new in the data domain AI with respect to Ashish’s research
[26:54] Most rewarding professional experience
[29:50] Next steps

Notable Quotes

Data science is the kind of thing we have to learn that kind of technology every day so we can just solve the business problem quickly. So we have to update ourselves on a day-to-day basis.
We always have to see how we visualize what is happening behind the algorithm to understand which feature is important.
We have to create a bridge that makes us unique from other data scientists, so it will just add up new research and technology.
Every data scientist has to research because they’ll know a lot of things, get a lot of experience by doing that, and learn new things from that.
So as a chief data scientist, my role is always to focus on the business; based on that business, I will prepare the architecture.
Aligning business objectives to an evolving AI/ML project takes a lot of time, but we have to do everything quickly as a fast-growing company.
I always prefer the data scientists always go where the client is, and then he or she can directly ask the question to the business owner or the subject matter expert where they want to design something. Communication is important from that perspective to understand the business and the requirements.
Transforming the business into a data science problem is more important than designing an AI/ML system.
Once you understand the business, you understand the KPI, you understand which business the client is focusing on or on which model the client is focusing. So that is a bridge.
Never go behind money; go behind your instincts and interest. Because when interest drives you, you will get success.
To become a great data scientist is unstoppable!