There's a Lot We Don't Know About Ebola

If you squint a lot, that is one approach that is being pursued. ZMapp is designed to bind to Ebola virus particles, preventing them from binding to cells.

https://en.wikipedia.org/wiki/ZMapp

That sort of system would be a pretty lofty goal, but realize that all of the challenges here are biological or biophysical. How do you recognize a pathogen? How do you target and deliver to specific tissues? How do you avoid the host immune system and degradation? Even if the nano-scale tech were there for the computational aspect, the real challenges would be the same ones scientists currently face, except on a larger scale.

That's sort of not too far off from how the human immune system works. The internal database would be your collection of antibodies[1] which match epitopes[2] presented by MHCs[3] on cell surfaces. The presence of Pathogen Associated Molecular-Pattern[4] should trigger a response. In terms of adding to the database, I think monoclonal antibodies[5] have been used to that end. I'm not entirely up-to-date and this is very simplified, so take with a grain of salt.

[1] https://en.wikipedia.org/wiki/Antibody

[2] https://en.wikipedia.org/wiki/Epitope

[3] https://en.wikipedia.org/wiki/Major_histocompatibility_compl...

[4] https://en.wikipedia.org/wiki/Pathogen-associated_molecular_...

[5] https://en.wikipedia.org/wiki/Monoclonal_antibody#Therapeuti...

Is called "leucocytes", we program it with vaccines since many years ago, and they work fine in most of the cases.