Human-Powered AI · Drug Discovery

Develop drugs
differently. Faster.

We fuse two decades of wet-lab drug-hunting experience with proprietary machine learning — turning real empirical data into clinical candidates. Not theory. Evidence.

Start a Program → See how it works

6 movs. 5 years

—90%Development time

5×Probability of success

2M+Empirical data points

The problem with theory

The industry bets billions on educated guesses.

Traditional drug discovery is a theory-first funnel — target to hit to lead to trial. At every stage, promising candidates fail. The math is brutal.

Patients have unmet needs

Biologists identify potential targets

Chemists create potential solutions

Companies screen millions of compounds

Researchers test dozens of "hits"

A few candidates reach animal studies

A handful reach human clinical trials

FDA Approval Rate

Just 1 in 10,000

drugs that begin the traditional discovery process ever receive FDA approval.

MolecuLern was founded to improve these odds — replacing trial-and-error with candidates that already carry the fingerprints of real, drug-like success.

The MolecuLern difference

Accelerated AI discovery — without sacrificing the integrity of the process.

01 / The Human

Drug hunters, not just data.

Our co-founders have brought more than 20 drugs to the clinic, authored 150+ manuscripts, and hold over 100 patents. That hard-won intuition trained MolecuLern to think like a world-class human drug hunter — knowing which candidates are worth chasing and which look good only on paper.

02 / The Machine

Learning from real wet-lab evidence.

Where most AI trains on theory, MolecuLern learns from over two million real empirical data points — actual synthesized compounds and measured outcomes across decades of successes and failures. That grounding is what makes its predictions accurate, fast, and reliable.

The platform

A living chemical library of over five billion designs.

Curated from millions of compounds and supported by algorithms that identify 40+ key properties needed to select a lead — spanning three modalities.

NCEs

New Chemical Entities

Synthesized, drug-like NCEs14,000
Virtual NCE designs8.8M
Key properties modeled40+

Fragments

Fragment-Based Discovery

Synthesized fragments36,000
Real empirical data points2M+
Screenable designs · Synthesis on Demand5B+

Degraders

Targeted Protein Degradation

Novel PROTAC linkers & ligases200+
Novel molecular glues100+
Modalities supported3

−0%

Reduction in development time versus traditional discovery.

Time to candidate

0×

Increase in the probability of clinical success.

Higher-quality leads

6 mo

To a validated lead — a process that traditionally takes five years.

Speed to lead

Why it works

Three advantages, built in.

/ 01

Data Library

Nearly 14,000 drug-like NCEs, 8.8M virtual NCEs, 36,000 synthesized fragments, plus PROTACs and molecular glues — all backed by algorithms that rapidly surface the 40+ properties that define a real lead.

/ 02

Dialed-in Experts

Founders who have brought 20+ drugs to the clinic, published 150+ works, hold 100+ patents, and founded 10 companies with multiple exits. That expertise trained the platform to think like a human drug hunter.

/ 03

Direct to Companies

Built to flexibly meet your need — whether you want to optimize an existing compound, develop a new drug from scratch, or screen our proprietary library. MolecuLern is ready to collaborate.

The humans driving innovation

Meet the drug hunters behind the engine.

The product of over two decades of work — pioneers among the first to combine computational techniques with wet-lab "fingerprints" for drug discovery.

Dave Bearss

Co-Founder

Two decades in small-molecule drug development, with a track record of translating computational insight into real clinical candidates and successful company exits.

Hari Vankayalapati

Co-Founder

A pioneer in medicinal chemistry and structure-based design, blending deep chemical intuition with modern machine learning to hunt drugs at scale.

100+Patents

20INDs filed

20+Drugs to clinic

150+Publications

10Companies founded

Ways to leverage MolecuLern

Wherever you are, we can collaborate.

01Hit Identification & Lead OptimizationFind and refine the compounds worth chasing.

02End-to-End Full CollaborationA complete discovery program, run together.

03Custom Compound OptimizationImprove the drug-like properties of your molecule.

04Custom Library ScreeningScreen your library against our models.

05Screen Our Physical & Virtual LibraryAccess 5B+ designs across three modalities.

Develop drugsdifferently. Faster.