Great question — answering as the UphillTech team:
A machine learning algorithm is a more or less an abstract set of rules, mathematical functions, or steps that are required to formally state how the computer can learn to make, e.g., predictions. "Model" typically implies that a training process for determining weights (unknown model parameter, e.g., the coefficients of a polynomial that is fitted to some training data) already was performed. Note, however, that in the literature or in the web notions are not always well defined or used consistently: there is probably consensus about "algorithm" but sometimes "model" is also used instead of "architecture" (in the deep learning community). Unfortunately, there are a few community-dependent...
Full answer by BanDoP: https://datascience.stackexchange.com/a/123988 (CC BY-SA 4.0).
Want it for your business? UphillTech.