ANDROMEDA 1 Maps 181 PLGA Formulations

This is a compelling example of AI making experimental search more efficient, not replacing laboratory validation.

• –The campaign optimized conflicting objectives including loading, release, viscosity, and injectability simultaneously
• –Intermediate-molecular-weight polymers delivered the best balance between sustained release and manageable viscosity
• –Four lead candidates emerged from 181 experiments, giving researchers multiple viable paths instead of one fragile optimum
• –The result depends on automated experimentation and domain-specific polymer expertise as much as the AI model
• –The work remains preclinical: in-vitro release results still need stability, scale-up, and in-vivo validation