AI for cancer research, built for real‑world impact
Model design
Purpose‑built networks for multi‑modal cancer data
Computational biology
ML that respects biology: signals, mechanisms, and noise
Clinical translation
From lab demos to tools that fit real clinical workflows
Our Approach
Building the backbone for computational oncology
We focus on the unglamorous but essential pieces: clean data, robust methods, careful evaluation. The goal is simple: trustworthy systems that help patients and clinicians.
Foundational research
Multi‑scale representations that connect molecules to outcomes
Algorithm development
Methods tuned for messy, shifting, high‑dimensional oncologic data
Clinical integration
Designed for deployment: privacy, speed, and workflow fit
Research Focus
Computational methods that matter
Where advanced computation can actually move the needle
Early detection systems
Combine signals across modalities to catch disease sooner
Tumor microenvironment modeling
Model cell‑cell interplay to anticipate response
Treatment response prediction
Forecast outcomes to help choose better therapies
Advanced imaging analytics
Read subtle patterns that support precise diagnostics
Genomic intelligence
Decode genomics to reveal targets and biomarkers
Clinical trial optimization
Design smarter studies and stratify patients well
Core Technologies
Building foundation models for oncology
Not just off‑the‑shelf AI. We’re shaping frameworks built for oncology’s data and decisions
Efficient computing
Do more with less. Reduce compute without losing the biology
- Sparse attention for long genomic sequences
- Adaptive compression for edge‑friendly pilots
- Quantization that keeps biological signal intact
Explainable AI
Clinically useful means understandable. We prioritize clarity
- Multi‑scale attention maps you can trust
- Causal views of treatment effects
- Clear, structured reports, not black boxes
Learning from little data
Rare cancers aren’t rare to real people. Learning must be sample‑efficient
- Meta‑learning for rapid adaptation to new subtypes
- Contrastive learning for molecular similarity
- Synthetic data with biological guardrails
Open research infrastructure
We share tools, benchmarks, and models so progress compounds
- Modular multi‑modal integration
- Standardized, fair benchmarks
- Pre‑trained models for research use
Join Our Team
Help shape the future of cancer intelligence
We’re a small, focused team. If you like rigorous work, kind teammates, and hard problems, let’s talk.
Build with us
Great work comes from diverse, thoughtful teams. We keep the bar high and the egos low.
- Work on foundational problems that matter
- Collaborate with leading minds across ML and cancer biology
- Shape the roadmap of transformative cancer research
Machine learning scientists
Design new architectures for cancer data
Computational biologists
Bring mechanisms and data together
Software engineers
Scale and ship reliable systems
Clinical researchers
Guide what matters clinically