Systems being built across discovery, invention, mathematics, health, embodied intelligence, and AI-native research operations.
Each system combines foundation models with retrieval, tools, memory, coding, planning, evaluation, domain knowledge, and expert oversight to support real use in complex settings.
An active ecosystem of specialist agents for strategy, invention, science, policy, and domain advice, designed to orchestrate serious work rather than single prompts.
Status: active internal build environment across multiple research threads.
A discovery system for hypotheses, constructive-science programs, rediscovery benchmarks, operator libraries, and evaluation of scientific claims.
Status: active build with testbeds spanning materials and the physics of intelligence.
Related platformA mathematical discovery and theorem handoff stack connecting conjecture generation, proof planning, Lean formalization, and draft auditing.
Status: active bridge between scientific reasoning and formal proof workflows.
A lightweight invention-management environment exploring energy-aware AI and evidence-centric MRI report generation with provenance-bearing intermediate states.
Status: active prototype environment with a live MRI research direction.
Related platformA starter architecture for intelligence built around vision, audition, touch, proprioception, world modelling, memory, planning, action, and homeostatic safety.
Status: active architectural prototype connected to first-principles intelligence work.
Systems for supervision, literature search, ideation, experiment planning, reviewer simulation, and grant strategy in an era where AI is part of the research team.
Status: active internal infrastructure for how research groups may operate next.