Automating 2D data for best understanding about likely activity and structure for prioritization of screening

We have tens of thousands of novel but obtainable and researchable small molecules from a single natural microbiome (see for Biosortia technology article).  We want to exploit as much of the easy-to-create 2D analytical data to prioritize and rank these existing (and working in a microbiome) small molecules.  The workflow could be in excess of 50,000 novel small molecules in a single microbiome; we will have complete genomic information.  We are experts at Industrial Scale Microbiome Mining, which allows us to obtain this unique resource at the quantity and quality needed for a complete investigation, including 3D structure, but utilizing a priority workflow along with screening will focus efforts.  We are looking for experts in computational analysis, ML, and AI; thus we certainly see ChemDraw as one of the tools of understanding; we need guidance on the workflow of this volume to reach the goals laid out.  RY