Building structured memory for AI agents: MindGraph (mindgraph.cloud), a typed knowledge graphs + hybrid retrieval, so agents remember more than the last few turns. Runs on mnestic (mnesticdb.com), a Rust DB forked from CozoDB and tuned for the workload. Also Thumos Care (thumoscare.com): a preventive-care platform over a medical knowledge graph with clinical guidelines. The throughline across everything I build is the same: graphs as the connective tissue for memory and reasoning. Off the clock: religious and philosophical texts (Zen, Confucius, reading the Bible as parable), geopolitics, and the consciousness/neuroscience question of what memory even is. NYC, often remote. Happy to talk knowledge graphs, agent memory, or any of the above.