Learn Data Science With Mukeshram

When I started working with Inference, I quickly realized there was a gap between theory and what actually happens in practice. This post is about how i optimised a slow ml inference api. I'll walk you through what I learned, what tripped me up, and the lessons that stuck with me. No fluff — just honest notes from someone who went through it. Introduction to Optimizing a Slow ML Inference API I still recall the frustration of dealing with a slow ML inference API. The latency was unbearable, and it seemed like no matter what I did, I just couldn't get the performance I needed. But after weeks of trial and error, I finally managed to optimize the API and achieve significant improvements. In this article, I'll share my experience, the mistakes I made, and the lessons I learned along the way. The Initial Challenges When I first started working on the ML inference API, I was excited to see it in action. However, my enthusiasm was short-lived. The API was slow, and the laten...

When I started working with Monitoring, I quickly realized there was a gap between theory and what actually happens in practice. This post is about why model monitoring matters more than model training. I'll walk you through what I learned, what tripped me up, and the lessons that stuck with me. No fluff — just honest notes from someone who went through it. Introduction to the Importance of Model Monitoring As I delved into the world of machine learning operations (MLOps), I quickly learned that the real challenge lies not in training a model, but in ensuring it continues to perform well after deployment. I've seen firsthand how a model that excels in testing can silently deteriorate in production, often due to data drift - a change in the distribution of input data that can render a model ineffective. My experiences have taught me that model monitoring is not just a nice-to-have, but a critical component of any successful MLOps strategy. The Dangers of Assuming Deployment ...