From particle physics to foundation models,
our researchers have lived at the edge.
Labs puts them inside your team.
Research programs
Our research programs span scientific ML, frontier model engineering, and embodied AI: built for teams who need more than a fine-tuned model.
Frontier model engineering
We design research programs around models that reason across language, documents, imagery, and structured context, with evaluation tied to real business value and scientific tasks.
Vision-language pipelines for document, scene, and workflow understanding
Long-context reasoning systems with rigorous benchmark design
Training, adaptation, and deployment strategies built around task fidelity
Scientific machine learning
We apply modern ML to simulation-heavy domains, blending data-driven models with physical structure to speed up inference, improve approximation quality, and unlock new tooling for research teams.
Surrogate models for expensive simulation workloads
Hybrid pipelines combining numerical priors with neural architectures
Experiment loops tuned for robustness, calibration, and interpretability
Robotics research
We research embodied intelligence for autonomous land and aerial platforms, focusing on decision-making under uncertainty, multi-agent coordination, and operation in contested or communication-denied environments.
Collaborative mission planning and multi-agent coordination without reliable comms
Autonomous navigation and spatial reasoning in GPS/radio-denied areas
High-autonomy decision architectures for embodied systems in unstructured environments
We work best on problems where multimodal reasoning, scientific modeling, autonomous systems, or difficult document understanding need a team that can think like builders and researchers at the same time.
