01.1
Generate
Generative protein and binder design.
AI × computational protein engineering
Designing proteins for the real world.
Generative protein design and physics-informed evaluation across molecular stability, structural integrity, aggregation and developability.
Supported through hessian.AI · Lean AI Startup Funding (LAISF) 2026/27
01 / Scientific position
From generation to reliability.
Protein design does not end when a sequence is generated. PRIB combines generative models with physics-informed evaluation to identify candidates more likely to remain stable, structurally coherent and experimentally viable.
The objective is not to replace experimental validation. It is to make earlier design and prioritization decisions more informative before expensive experimental work begins.
01.2
Evaluate
Physics-informed assessment of stability and structural integrity.
01.3
Prioritize
Multi-signal ranking across aggregation, developability and experimental risk.
01.4
Learn
Use experimental outcomes and failure signals to improve future design decisions.
02 / Engineering approach
A closed design loop,
grounded in physical behavior.
Design space
Generate candidates and representations suitable for computational screening.
Physical signals
Interrogate stability, topology, conformational risk and aggregation-relevant features.
Decision layer
Integrate heterogeneous evidence into candidate ranking and interpretable prioritization.
Experimental feedback
Use outcomes and observed failure modes to refine future design and evaluation decisions.
03 / Selected signals
Evidence over adjectives.
ICLR 2026 GEM × Adaptyv RBX1
7 / 21
submitted designs selected for experimental testing.
Built through a five-day protein-design workflow // competition scale >12,000 submissions
Workflow build
5 days
from challenge entry to an operational design workflow.
Competition scale
12k+
submitted designs across the RBX1 challenge.
04 / Selected milestones
Active research. External signal.
Anthropic AI for Science supportSupport for PRIB's AI-for-science research program.
Festival of Biologics BaselInvited scientific presentation on molecular fragility and downstream robustness.
Four NeurIPS 2026 workshop papers acceptedAcross AgenticLS, ESR, PhysWorldAI and Robotics World Modeling.
TU Darmstadt IdeenwettbewerbSelected as a finalist in the research category.
05 / Research
Accepted work at NeurIPS 2026.
Selected team research accepted across NeurIPS 2026 workshops, spanning protein reliability, embodied reasoning, physical AI and robot world models.
Workshop papers · 2026Accepted
Learning What to Measure Next: Failure-Aware Closed-Loop Evidence Allocation for Protein Reliability
NeurIPS 2026 · AgenticLS
R01
Accepted
Scale2Real: Dimensionless Physical Representations for Cross-Scale Humanoid Locomotion
NeurIPS 2026 · 2nd Embodied Spatial Reasoning Workshop (ESR)
R02
Accepted
Prediction Fidelity Is Not Decision Fidelity: A Counterfactual Action-Ranking Diagnostic for Robot World Models
NeurIPS 2026 · Robotics World Modeling
R03
Accepted
Dynamic Similarity for Physical AI: Learning Locomotion Across Physical Scales
NeurIPS 2026 · PhysWorldAI
R04
Acceptance refers to NeurIPS 2026 workshops, not the NeurIPS main-conference proceedings.
06 / Institutional support
Built inside a public AI research ecosystem.
PRIB Labs is supported through the Lean AI Startup Funding (LAISF) 2026/27 program of hessian.AI — the Hessian Center for Artificial Intelligence.
The program supports research-driven AI teams in validating technology, developing prototypes and building a credible path from scientific work to application.
Research ecosystem
hessian.AI
The Hessian Center for Artificial Intelligence brings together AI expertise across 13 universities in Hesse, spanning research, infrastructure, transfer and entrepreneurship.
Program
Lean AI Startup Funding
LAISF 2026/27 is a six-month scholarship program for innovative AI startup teams, focused on concept validation, prototyping and business-case development.
Venture ecosystem
AI Startup Rising
hessian.AI's startup initiative connects research-driven AI teams with funding, infrastructure, mentors, industry and Germany's broader deep-tech ecosystem.
Public support context
hessian.AI is funded by the Hessian Ministry of Science, Research, Art and Culture. AI Startup Rising is supported by the German Federal Ministry for Economic Affairs and Energy (BMWE) and EXIST.
07 / Core team
Physics, computation,
and translation.
Founder · Scientific Lead
Dr. Haipei Liu
Protein mechanics, single-molecule biophysics, stability frameworks and computational workflows.
Co-Founder · Physics-informed AI
Dr. Xianglong Peng
Deep learning, computational mechanics, energy-based modeling and perturbation-aware algorithms.
Co-Founder · Business Development
Bingyi Zhao
Industry translation, product strategy, partnerships and market-facing decision logic.
08 / Contact
Work with PRIB.
Scientific collaboration, protein engineering, industrial research and early partnership discussions.