One of my goals as an academic is to facilitate collaborations between my field ("pure algebra") and other sciences. We can't expect administrative staff (for instance, the Dean of Science) to help much with this. So it's our responsibility, as scientists, to find alternatives.
(Q1) Would it make sense to add a MathOverflow feature with this in mind? The intent would be to give references to papers, list a few "open problems" in the field; and then find collaborators. MathOverflow could add a "tag" saying "applied algebra project"; and those interested could reply to the question with their specific projects that they'd like to collaborate on.
(Q2) Are there any websites which would be useful for finding collaborators? Poly-Math isn't ideal either; it's intended for very specific projects.
I've listed some topics below.
(#1) Topological quantum computing (uses Jones polynomial and braid groups)
(#2) Geometric complexity theory (approach to P=NP that uses algebraic geometry)
(#3) Algebraic data science (approaches to data science that use algebra - for instance, category theory; or persistent homology)
(#4) Algebraic statistics (for instance, Sturmfels-Pachter "Algebraic statistics for computational biology")