KiraLOGIC: AI That Learns From Real Biology
AI trained on real function, not static readouts
KiraLOGIC is our AI-powered platform that designs multiplex edits to keep immune cells functional under tumor microenvironment (TME) pressure.
It learns directly from real phenotypes — cytotoxicity, cytokine release, RNA-seq, and persistence in 3D tumor models — so our designs reflect how cells actually behave, not just what they express.
How It Works
‘Intervention’ includes gene edit combinations applied to cells.
Outputs shown are representative, not exhaustive
-
Design — Generate candidate edit sets matched to each tumor’s dominant suppression modes.
-
Test — Run multiplex assays in suppression-rich models (spheroids, organoids, xenografts).
-
Learn & Select — The model ranks combinations for durability and nominates the next in vivo candidates.
Why It Matters
-
Fewer cycles → compress discovery time
-
Better combinations → multi-pathway protection by design
-
Scales broadly → applies across T, NK, and γδ T cells
The KiraLOGIC Flywheel
Smarter with every dataset. Stronger with every program. Every experiment trains KiraLOGIC—powering smarter, more durable cell therapies.
Foundation
GBM Program validates TME resistance logic
Acceleration
Accelerates application across tumor types
Scale
The same protection scales across cell types
Moat
Every dataset sharpens predictions & defensibility
