TensorFlow is a free, open-source python library originally created by Google for building machine learning pipelines. TensorFlow bundles a whole series of models and algorithms for different purposes, mainly with a focus on (deep) neural networks. These neural networks are trained with training data and then produce a model that can be used for inference or to make predictions. The python interface is used to provide a convenient front-end API for building applications with the framework, which are then run in high-performance C++. Tensorflow supports large-scale projects, and can therefore work with enormous amounts of data.
What does a TensorFlow developer do?
Developers or Data Scientists use TensorFlow to create dataflow graphs. These are structures that describe how the data (tensors, or multidimensional data arrays) flow through different processing nodes (basically mathematical operations). Hence the name TensorFlow.