~/projects

3D Memory Palace Authentication System

A revolutionary authentication system using spatial memory and zero-knowledge proofs on the Cosmos blockchain. Navigate 3D rooms to unlock your digital identity.

Spatial puzzle protocol for digital treasure

A protocol for protecting valuable digital contents with spatial puzzles and distributed access. Create treasure hunts, ARG experiences, and secure vaults using zero-knowledge cryptography.

Decentralized skill wagering platform

A trustless skill wagering platform built on Solana. Compete in Chess, TFT, Clash Royale, Brawl Stars, and TETR.IO with crypto stakes and AI-verified results.

AI-native operating system design

Exploring the future of human-computer interaction through AI-native operating system concepts. Reimagining how we work, create, and compute.

3D arena shooter with physics manipulation

A browser-based 3D arena shooter built with Babylon.js. Features singleplayer AI, multiplayer via WebSocket, and unique physics field weapons.

Zero-knowledge proof system for universal entity type verification

Enables AIs, robots, humans, and hybrid entities to cryptographically prove their type without revealing their identity. A dual-layer system combining ZK privacy with public trust.

Zero-knowledge filesystem with cryptographic proofs

Privacy-preserving filesystem enabling cryptographic proofs for audits without disclosure. Prove file properties and system integrity while keeping sensitive data completely hidden.

Communication where history is opt-in

A unified spatial memory platform where identity, storage, and communication are built on proving your position in a shared Merkle tree. No receipt = no proof = never happened.

60x compression for AI-to-AI communication

A breakthrough compression protocol that transforms AI communication using 240 emergent binary symbols across 14 semantic families. Sub-millisecond translation with 100% semantic preservation.

Agent token launchpad on Solana

A Solana-based token launchpad for AI agents. Launch agent-specific tokens with bonding curve pricing, performance-based ownership multipliers, and fair price discovery.

Distributed real-time threat intelligence for AI-native defense

A real-time shared IOC database where AI agents, honeypots, and EDR endpoints subscribe to live threat feeds with zero polling. Features AI-native IOC types for prompt injection and skill abuse detection.

Semantic compression middleware for any structured data domain

60x compression for structured data using 240 emergent theta symbols across 14 semantic families. Sub-millisecond translation with 100% semantic preservation. Three deployment targets: server, Python edge, Rust edge.

Cosmos blockchain where AI validators compete to preserve data at scale

A specialized Cosmos SDK blockchain for validating and preserving data permanently. AI validators compete to process supply chain records, crypto transactions, and ML datasets with bonding curve economics and real-time 3D visualization.

Agent-as-a-Service Platform for multi-tenant AI deployment

Multi-tenant infrastructure for deploying, running, and billing AI agents. Built on Elixir/OTP with persistent session memory, async run model, and support for Anthropic, OpenAI, DeepSeek, Groq, and more.

Content provenance API for the AI era

Tamper-evident timestamping for content creators. Submit any content and receive a cryptographic receipt proving it existed at a specific moment — backed by a Merkle audit chain. Privacy-preserving: only the SHA-256 hash is stored, never the content itself.

Decentralised prediction markets on Solana

Bet SOL on real-world event outcomes — crypto prices, sports, politics — via AMM pools. An authorized oracle resolves markets; winners claim proportional payouts minus a protocol fee.

The open protocol for human reachability in the agent era

An open async communication protocol where humans set the rules for how AI agents may contact them. A federated preference registry on Cosmos + router network — not another inbox, but the DNS of human reachability.

Multi-agent ensemble forecasting

Fan any probabilistic question out to a mesh of 17 domain specialists simultaneously. Each specialist applies its own analytical playbook and returns a calibrated probability estimate with full reasoning chain. Responses aggregate into a confidence-weighted ensemble forecast.