Work
01
ATOM
Verified reasoning through multi-model consensus
A reasoning engine that decomposes complex questions into atomic claims, then verifies each through multi-model consensus rather than trusting any single LLM.
View project02
Lumin
A sanctuary for mastery in an age of AI
A learning platform that inverts the default: AI produces draft knowledge artefacts, you curate and refine them into genuine understanding.
View projectA retention system built around the forgetting curve, scheduling reviews at the precise intervals where memory is about to decay.
Also exploring
SOCI Social cognition in multi-agent systems
04