About VibeMesh Labs

What We Do

VibeMesh Labs is a computational research company focused on predicting antibiotic drug synergies using machine learning. We develop models that identify combination therapies against drug-resistant pathogens — starting with carbapenem-resistant Acinetobacter baumannii (CRAB), a WHO critical-priority threat.

Our approach combines resistance mechanism modeling, molecular fingerprint analysis, and efflux pump substrate profiling to predict which drug pairs will show synergistic activity. We validate predictions through independent contract research partnerships.

What We've Found

Our model predicted azithromycin + meropenem as a top synergy candidate against carbapenem-resistant Gram-negatives — a prediction independently confirmed by clinical observations from a collaborating research group treating patients in active care.

We are currently pursuing experimental validation of multiple drug combinations through CRO partnerships, with checkerboard synergy assays, biofilm disruption testing, and time-kill kinetics.

Founder

Our Approach

We believe the next generation of antibiotic therapies won't come from new molecules alone — they'll come from smarter combinations of drugs we already have. Our ML pipeline screens thousands of potential pairings to identify the ones most likely to show synergistic activity, then validates those predictions with independent laboratory and clinical data.

• Computational prediction — Gradient boosting models trained on published synergy datasets
• Clinical correlation — Partnering with researchers who have real-world treatment data
• Experimental validation — Independent CRO testing to confirm or refute predictions
• Open science — Publishing findings and methods transparently

The Blog

Our blog covers our antimicrobial resistance research, methods, and findings, along with occasional mathematical essays exploring unexpected connections in pure and applied math.

VibeMesh Labs is independent and self-funded. We publish openly because science should be visible.