BEAM LAB
Benchmarking all-atom biomolecular structure prediction with FoldBench
Nature Communications 2025

Benchmarking all-atom biomolecular structure prediction with FoldBench

Cryo-EM reveals mechanisms of natural RNA multivalency
Science 2025

Cryo-EM reveals mechanisms of natural RNA multivalency

Fast, sensitive detection of protein homologs using deep dense retrieval
Nature Biotechnology 2025

Fast, sensitive detection of protein homologs using deep dense retrieval

π-PrimeNovo: accurate non-autoregressive deep learning for de novo peptide sequencing
Nature Communications 2025

π-PrimeNovo: accurate non-autoregressive deep learning for de novo peptide sequencing

Research

We focus on AI for science.

Publications →
Focus

Agentic systems for scientific discovery

Autonomous reasoning and workflow automation for evidence synthesis, target discovery, and experimental design.

Agentic systems for scientific discovery
Focus

Foundation models for drug discovery

Developing foundation models that connect sequence, structure, and function across genomes, proteins and small molecular.

Foundation models for drug discovery
About Us

BEAM LAB

BEAM Lab develops AI methods for scientific discovery, with emphasis on foundation models connecting biomolecular sequence, structure, and function.

  • Biomolecular modeling: proteins, antibodies, and RNA

  • Structure prediction/design, proteomics, and AI-driven drug discovery

  • Agentic science for autonomous reasoning and research workflows

Join us

We are recruiting.

PhD students, research interns, and full-time researchers are welcome to apply.

Open positions