ML Engineers’ day-to-day consists of data-centric problem-solving. To do so, they have to face quite a few challenges such as deploying models to production at scale, collaboration among data scientist teams, and fetching large volumes of data from remote, dynamic environments.

In order to help ML Engineers to make sense of these complexities in an easy and intuitive way, the next generation of MLops tools will have to embrace these new challenges and adapt accordingly.

In this talk, we will discuss the challenges of creating an ML-friendly product. I will share insights on creating clear, intuitive, easy-to-use tools specially made for ML engineers facing these daily challenges. We will deep dive into the principles of creating such tools with real-life examples from my experience as the lead product manager at Dagshub and a former software engineer. Finally, we will list practical tips and best practices such as visualization and data display options that will be useful for creating any ML-friendly tool.