Across 50+ XR projects and prototypes, the same problem kept showing up: spatial systems rarely carry useful context from one session to the next. My work uses XR and AI to make complex workflows easier to understand, adapt to, and act on while the research is still in progress.
OpenSpatialAI is the early GitHub/code surface for APIs, memory, and adaptive XR behavior patterns. Harmony remains the research frame behind it.
Canonical personal profile and research record. Public experiments live at Varun Innovates; scoped client delivery lives at VeeRuby. Same work, different settings: research, prototypes, and practical systems.
Harmony investigates three interconnected problems: how XR systems perceive continuous human state, how persistent spatial memory enables adaptation across sessions, and how explainable AI mediation builds trust. The point is not to overstate the work early; the point is to make the research problem, artifacts, and limits inspectable.
External Proof Signals