About

I'm Shan Rizvi, a context architect building AI systems that model human cognition through knowledge graphs, semantic memory, and psychological profiling. My work gives language models the ability to remember, reason, and respond with relationship awareness.

Recent projects span commercial AI and open-source infrastructure: Thumos Care, a health companion that models complete medical history for safer, more personalized guidance; mindgraph-rs, a production-grade semantic memory graph in Rust with 48 node types and 70 edge types for agentic systems at scale; Engram, a neuroscience-inspired MCP server giving AI assistants long-term relational recall; a 32-billion parameter scientific reasoning engine trained on landmark papers through graph-powered reinforcement learning; and MatSyn, an agentic knowledge graph for compound synthesis that won second prize at AGI House's AI4Science Hackathon.

I am driven by the conviction that harmonization of consciousness has been humanity's highest ambition across diverse cultures for thousands of years.

My eclectic pursuit of knowledge allows me to see bridges between neuroscience, mystical traditions, AI systems engineering, and geopolitical strategy. These aren't separate domains—they're different symbolic systems concerning the same underlying reality. What mystics intuited about consciousness, neuroscience now measures, and AI can operationalize.