Research
Eight areas, one discipline.
Our work spans the stack of a modern AI system, from the models to the data and infrastructure beneath them. What ties it together is a shared standard of evidence.
Foundation models and language systems
How large pre-trained models are adapted, composed, and evaluated as components of larger systems, and what they can and cannot be trusted to do.
Explore 02Explainability and interpretability
Methods that make model behaviour legible to the people who depend on it, from feature attribution to mechanistic analysis, and the evidence that an explanation is faithful.
Explore 03Multi-agent and agentic systems
Architectures in which several model-driven agents plan, delegate, negotiate, and act, and the coordination, safety, and cost questions that come with them.
Explore 04Safety, alignment, and evaluation
Making sure systems do what their operators intend, fail in ways that are detectable, and are measured by benchmarks that cannot be gamed by accident.
Explore 05Data science and applied machine learning
The methodology of turning fragmented data into tested predictions, with baselines first, uncertainty shown, and every result reproducible from raw sources.
Explore 06Big data, analytics, and ML systems engineering
The infrastructure beneath the models, from data lakes and warehouses to serving, CI/CD, and observability, built so that research results can become dependable systems.
Explore 07Spatial and geospatial intelligence
Learning from satellites, land cover, census, infrastructure, and local records on a common spatial index, to see where places are changing before the change is obvious.
Explore 08Forecasting and decision intelligence
Models that forecast change over years, not seconds, and the decision frameworks that turn a probabilistic forecast into an action with a known risk.
Explore