Forecast biomass with a dynamic data set

We're looking to automate the parametrization of a series of forecast models for farm animals. Our data set is dynamic, and we want to code a neural network forecast that keeps updating its weights/parameters (is retrained) each time new data is entered. The code must be developed in Python.
You will work with a field specialist to develop the models, select the inputs and validate the outputs, and with our CTO to implement the solution into the project.