Cost-effective fast document filtration for legal corpora
Leveraging classical ML and NLP systems, we classify legal data as relevant or irrelevant to court filings without dense encoders or computationally intensive architectures.
Download paper
Applied intelligence research
Explore with usResearching how intelligence
becomes system
Our research looks at how probabilistic AI can be structured into more reliable, coordinated and increasingly autonomous systems. From deterministic control layers to multi-model collaboration and agentic workflows, we explore the architectures that can move AI from isolated outputs to working systems.
Deterministic systems
Research into the mathematical, algorithmic and structural foundations that govern system behaviour, including probabilistic modelling, simulation, optimisation, boundary testing and empirical validation.
LLM swarms
Exploring how multiple models and specialised agents can coordinate, distribute tasks, cross-check outputs and collectively solve problems beyond the capability of a single model.
Agentic systems
Studying autonomous and semi-autonomous systems that can reason, plan, use tools, act across environments and adapt through multi-step workflows under defined constraints.
Experimental tools
A growing collection of utilities designed for immediate use. We provide tools that anyone can easily plug in their own solutions or build directly on top of existing workflows.
Latest paper
