This post is drawn from the conclusion section of my thesis (User-Centric Machine Learning Systems). In this post, I tried to reflect on these questions:

What kind of ML systems problems are truly worth our time? How do we identify impactful and promising directions? How should we approach solving them thoughtfully?

Additionally, I share my perspective on where AI is headed and what the future of ML systems might look like. Hoping they are helpful for future researchers walking a similar pah.