I have been working as a software engineer for 15 years, and over the last 2 years, I have had the opportunity to delve into the world of machine learning — training models, building models, and deploying models. Currently, I am involved in all phases of the process, from ideation to go-live.
While there are thousands of articles and resources available that discuss the tools required to build machine learning models, only a handful of sources offer insights on how to serve these models effectively. I understand that some may argue this information is scarce; however, I believe that sharing my experience and knowledge in this area can contribute to the ongoing conversation and help others navigate the challenges of model deployment more easily.
Focus on the journey, not the destination. Joy is found not in finishing an activity but in doing it.
When building a product, creating it is just the first step; delivering it to the end-users is where the real challenge lies. This is especially true for ML models. Imagine a brilliant data scientist and their team develop an exceptional model, but if no one uses it, what’s the point? Building an ML model is difficult, but deploying it efficiently in production is equally, if not more, challenging. Now, let’s dive back into the technical aspects. 😁
Next topic will be on Feature Store
