Career Insight from an Amazon.com Machine Learning Scientist
I recently invited Dr. Hyokun Yun, machine learning scientist at Amazon.com to my monthly LinkedIn Live session featuring the LinkedIn Learning course: Data Science & Analytics Career Paths & Certifications.
Due to the time limit, we could not cover all the prepared questions for Dr. Yun, and one of them was:
As a seasoned professional in this area can you describe what a typical day might look like for a data scientist?
Below is his response.
- Each project tends to have multiple phases, and my day-to-day work changes a lot depending on which phase I am in.
- In the initial phase of the project, I collect product requirements from business owners & product managers, and engineers. I collaborate with engineers to develop a software system design that meets these product requirements.
- A good fraction of my time constantly goes to literature reviews. Machine Learning (ML) is still fast moving, and often a new technique enables what’s hitherto not considered feasible.
- I collect data, train models, evaluate their performance, and iterate until I reach the point the model is worth putting into production. Often these involve heavy software engineering, and I collaborate with software engineers.
- I then collaborate with software engineers to push the new model into production. I am responsible for developing an ML part of the code, so I write code and have it reviewed by engineers. I also review engineers’ code such that they are aligned with what’s needed for science.
- As I am becoming more senior, I have been spending an increasing amount of time on meta-level tasks: rather than doing these all on my own, creating a plan for executing these tasks and high-level technical directions of how they should be executed, etc.
For your information, below are links to the three LinkedIn Live sessions we had so far.
Wish there’s a link to watch past sessions... I’m simply too busy to commit to your live programs 😩