AI-guided protein design
Constrained design workflows for VHH-format binders, conditional protein switches, protease-sensitive linkers, and receptor-targeted delivery architectures.
COMPUTATIONAL MOLECULAR ENGINEERING
Intelligence at the
molecular scale.
We bring generative AI, structural modeling, and evolutionary optimization together to explore the design of proteins and molecular systems.
Explore our researchFrom computational hypotheses toward experimentally testable designs.
THE DESIGN SPACE
SEQUENCE → STRUCTURE → FUNCTION
AI +
BIOLOGY
GenerateModelEvaluateRefineBIOLOGY
Iterative search across competing design objectives
Conceptual framework · not experimental data01 / RESEARCH
Engineering possibilities.
Testing assumptions.
Our work centers on computational design and evaluation. These research directions describe workflows under development, rather than validated therapeutic products.
Structural bioinformatics
Protein-structure prediction and PyRosetta modeling support evaluation of predicted fold confidence, geometry, motif accessibility, interface packing, and local mutation sensitivity.
AI & optimization systems
Multi-objective evolutionary search and parallel scoring help compare candidate designs across structural and biophysical criteria, with Pareto-front management to expose tradeoffs.
02 / THE PLATFORM
A traceable path
from idea to candidate.
Scientific search needs more than a model. Our event-driven campaign framework connects design, evaluation, and candidate management while preserving the history of each experiment.
- Plug-in evaluators and parallel scoring
- Checkpoints and failure recovery
- Experiment provenance and technical documentation
- Pareto-front candidate management
Define constraints
Specify the design problem, geometric requirements, and competing objectives.
Generate & model
Explore candidates with generative methods, structure prediction, and molecular modeling.
Score & compare
Evaluate packing, accessibility, charge, and sensitivity alongside predicted confidence.
Plan experimental validation
Use computational evidence to inform staged wet-lab validation planning with academic collaborators.
03 / SCIENTIFIC APPROACH
Evidence before claims.
Computational scores help prioritize experiments. They do not establish binding, biological function, safety, or clinical efficacy.
Keep predictions distinct
Predicted structure confidence and biophysical scores are model outputs. Experimental measurements are a separate source of evidence.
Make tradeoffs visible
Multi-objective evaluation compares competing criteria without reducing every scientific question to a single score.
Preserve provenance
Campaign checkpoints, experiment histories, and documentation support inspection and recovery throughout computational work.
INNOVATIVE MACHINES
Research across AI and biology.
Our scientific work combines molecular-engineering strategy, computational platform development, partner coordination, and planning for staged experimental validation.
Research descriptions on this site are based on the lead scientist’s supplied professional résumé. No publications, quantitative performance results, or completed wet-lab validation are asserted here.