Challenges of ML development
No matter the type of ML persona or the type of model artifacts they produce, the challenges are the same across the board. The following diagram captures the three main challenges around data, people, and tools:
Figure 9.6 – Challenges in ML life cycle management
There is a plethora of ML options to experiment with, but there are only a few people with the skills needed to command the landscape; there are still fewer people who are able to use these tools properly, and even then, their efforts are at the mercy of the quality and completeness of the datasets available to them. Let us examine these challenges in more detail:
- The ML ecosystem is a rich one, with new tools, libraries, and frameworks mushrooming every day. There is a need to be able to experiment with them to check their efficacy without it being too disruptive and distracting. In other words, ML practitioners need the ability to layer in a new...
