Full-Stack + MLOps Engineer — Turn a Scientific Python Tool into a Multi-User Web Platform

We have a working AI platform for materials discovery for biologic delivery (formulation/excipient discovery for insulin, mAbs, peptides, vaccines). It's built in Python and currently runs as a local developer tool — an MCP server plus a partial FastAPI backend — on top of a heavy scientific-computing stack (OpenMM molecular dynamics, RDKit, Packmol, retrosynthesis, ADMET prediction). It works, and it's backed by published research. What it is not yet is a product: there's no login, no database, no per-user experiment tracking, and long simulations run synchronously on whatever machine launches them. Your job is to turn this codebase into a hosted, multi-user web platform where a user can sign up, log in, launch a discovery campaign, watch it run (on GPU when available), and come back later to a persistent record of their experiments, results, and audit trail. This is a real product with a real founder, not a throwaway prototype. We value clean, tested, documented work over speed. Here's the repo: https://github.com/otunmartins/biologix-platform Kindly check attachment for more info