8+ years building scalable web, mobile & backend · AI-powered products & agentic systems · Blockchain, DeFi & Solana
Full-stack, AI & Web3 — intelligent apps, agents, and on-chain products that scale.
🧠 Engineering OS — @truelogics
A platform that gives AI coding agents an organization's actual engineering decisions — its ADRs, rules and architecture — instead of general best practice. Built in Go, designed as a kernel with independent consumers.
| engineering-kernel | The kernel and the eng CLI — collect, parse, index, retrieve, classify. SQLite + FTS5, no embeddings, no vector DB, by design. |
| engineering-mcp | An MCP server that exposes that knowledge to Claude Code, so a review can cite the decision it contradicts. |
| engineering-review | The first consumer: pull-request review grounded in retrieved context rather than a diff in isolation. |
| engineering | The rulebook the platform reads — standards, ADRs, RFCs, and the policy governing what gets built. |
Developed through documented RFCs and controlled experiments — including one that measured why an AI client silently ignored a registered MCP server, and found the cause was deferred tool schemas rather than the tool descriptions everyone assumed.
Go · SQLite/FTS5 · Model Context Protocol · JSON-RPC · 9 RFCs
Open to new opportunities and interesting projects.