Production environments often differ from development environments in such a way that the additional development costs are needed to go from a proof of concept to an industrialized solution
that actually serves users can limit the feasibility of this transition.
Some wording suggestion : "Production environments often differ from development environments in such a way the feasibility of this transition can be limited by the additional development costs which are needed to go from a proof of concept to an industrialized solution that actually serves users."
To do so, they need to have access to substantial and diverse computing resources that enable them to tackle the volume and diversity of big data sources and leverage machine learning methods to better deal with these data
In line with my first remark : it seems to me that "the ability of statisticians to carry out innovative experiments" is not necessary linked with the fact that they handle "big data" thus my suggestion do drop the end of this sentence.
that actually serves users can limit the feasibility of this transition.
Some wording suggestion : "Production environments often differ from development environments in such a way the feasibility of this transition can be limited by the additional development costs which are needed to go from a proof of concept to an industrialized solution that actually serves users."
In line with my first remark : it seems to me that "the ability of statisticians to carry out innovative experiments" is not necessary linked with the fact that they handle "big data" thus my suggestion do drop the end of this sentence.