Member-only story
How to Rethink Your DS Stack š¤
In my previous posts, I have discussed how the data science tech stack has evolved over time and shared insights on how to operationalize machine learning with a unified approach.
I wish thereās a single tool that can solve all the pain-points for data science. Of course, thatās just wishful thinking.
As I begin building a machine learning team from scratch, I want to revisit how we evaluate the tools and stack available for machine learning and data science. I aim to assist new leaders, data scientists, and machine learning engineers in coming together to shape the way we work and make it more efficient.
Hamal Husain presented a talk on evaluating MLops tools at Stanford last year, which I found to be very informative and valuable. Today I would like to share some of the key insights from his talk.
Landscape
When you start from scratch, one of the challenges is being overwhelmed by the multitude of choices available. The landscape of tools and applications for data science and machine learning is vast and complex, as illustrated below. Trying to navigate through this landscape can be daunting, which is why it is important to have a clear understanding of the specific requirements and goals at hand.
