🚀 What I build: agentic-AI infrastructure — MCP servers, Agent-to-Agent (A2A JSON-RPC) integrations, Hedera Agent Kit plugins, on-chain agent identity via ERC-8004, and developer tooling that makes LLM systems shippable.
💡 Open to collaborate on: multi-agent orchestration, MCP tool design, A2A protocol work, agent identity, RAG systems, and platform-engineering for LLM apps.
⚡ Token optimization: designing prompt-caching, context-window discipline, and multi-host tooling that keeps agentic-AI systems cheap to run at scale.
🌱 Currently learning: long-horizon agent planning, eval-harness design, RAG retrieval quality, and Rust on the platform edges.
🛠️ How I work: 10+ yrs of platform / full-stack engineering. Senior IC who ships end-to-end — from schema design and protocol choices through production deploys and developer adoption.
💬 Ask me about: #AgenticAI #LLMSystems #MCP #A2A #HederaAgentKit #ERC8004 #PlatformEngineering #TypeScript #NodeJS #Rust
- Private MCP routing side project — designing a local MCP router that gives host apps one stable endpoint, routes tool/resource/prompt calls to downstream integrations, manages plugins, and benchmarks routed vs direct execution.
- Agentic AI infrastructure — building MCP servers, Agent-to-Agent (A2A JSON-RPC) workflows, and tool-calling integrations for developer-facing platforms.
- MCP / tool design — designing schema-driven tools with authentication, session handling, and test coverage for external developer use.
- Web3 agent identity — integrating Hedera Agent Kit and ERC-8004-style identity flows into AI-agent workflows.
- AI development workflow — maintaining a token-conscious multi-host local agent tooling setup across Claude Code, Codex, OpenCode, Cursor, and Gemini CLI.
Agentic AI / Agent Platforms
Languages
Backend / Platform
Cloud / DevOps
Data