🎓 eduNest
AI-native interactive education and automated courseware generation platform.
Next.js + Express + Supabase + multi-LLM + realtime voice + interactive H5 courseware.
🔎 LeadGen
Autonomous multi-tenant website generation and deployment infrastructure.
Lead discovery → LLM-generated UI → Vercel deployment → DNS provisioning in one automated pipeline.
Turning real Drosophila connectome data into executable neural circuits.
FlyEye is an open-source developer kit and interactive experiment built around a real, signed MaleCNS-derived Drosophila connectome subgraph.
3,376 neurons · 78,797 signed neural edges
JavaScript / TypeScript SDK · Python SDK · Hugging Face Dataset · Zenodo DOI · Interactive browser experiment · Reproducible graph extraction
Live Demo · GitHub · Dataset · DOI
A persistent agent world where autonomous agents perceive → retrieve memory → decide → act → receive environmental feedback → update memory.
Current work includes persistent multi-agent state, memory and relationships, social/economic interaction, governance, real-world geography and weather, external information, and explainable decision traces.
Agent infrastructure, execution, safety and governance.
Exploring how autonomous agents can be coordinated as reliable long-running systems with shared state, tools, memory and recovery mechanisms.
| Artifact | Project |
|---|---|
| Hugging Face Dataset | FlyEye Escape Neuron v1 |
| Zenodo DOI | 10.5281/zenodo.22829797 |
| npm packages | FlyEye graph / runtime / tooling |
| PyPI package | flyeye-graphs |
| GitHub Releases | Versioned FlyEye publication assets |
| Interactive demo | fly.fde.fan |
Agents · World Models · Simulation · Robotics · AI Infrastructure · Open Research
Building systems where AI can perceive, reason, act and continuously adapt to the world.