My understanding is that Google developed Tensorflow to provide a single pipeline between the data science model and a production system. The idea is to avoid a translation step between modeling the real world and implementing it in production.
Of course, most people don't have the resources of Google with a layer of data scientists and another layer of software engineers [and maybe a layer of data engineers in the mix too]. So the idea of a tool tailored to a small team's needs rather than those of Google seems like an interesting niche.
Leaf takes an imperative approach and explores an easier API (only Layers (Functions)[1] and Solvers (Optimizer Algorithms)), reusability through modularity and abstractions that keep the implementation and concepts to a minimum or rather abstractions that feel as familiar to a hacker as possible.
For future versions e.g., we want to explore what is practically possible with auto-differentiation via dual numbers and differentiable programming.
[1]: http://autumnai.com/leaf/book/deep-learning-glossary.html#La...