┌─────────────────────────────────────────────────────────────┐
│                                                             │
│     C A R L O   M I G U E L   D Y    ·   @carlomigueldy     │
│                                                             │
│     senior full-stack  /  ai-augmented product engineer     │
│                                                             │
└─────────────────────────────────────────────────────────────┘

I'm a senior full-stack product engineer with deep Web3 experience and a daily AI-augmented engineering practice. I'm not just a frontend, backend, or smart contract engineer — I'm the one who owns the feature from a Figma frame and a half-written spec all the way to a green CI run and a production deploy.

const miguel = {
  ownsEndToEnd: ['frontend', 'backend', 'smart contracts', 'tests', 'ci/cd', 'deploy'],
  shipsWith:    ['claude code', 'codex', 'mcp', 'sub-agents'],
  bestAt:       'turning vague product requirements into production systems',
  status:       'open to senior full-stack / web3 / ai-native roles',
};

🛰   MCP server platform — @AltLayer

Took over a backend handover of an MCP server exposing read/write tools across Uniswap, SushiSwap, Polymarket, 1inch, Balancer, OKX, DeBank, OpenSea and others. Reduced the surface from ~400 tools → ~200, then introduced a router-orchestration pattern that consolidated those into ~20 router tools while preserving correct tool selection from natural-language inputs. Wired in CDP server wallets so LLMs can perform on-chain actions — with mandatory human confirmation before any write. Built the chat UI (multi-model, OpenRouter-backed) showing tool calls inline.

🎯   Prediction markets platform — @Limitless Labs

Founding engineer. Architected the Nest.js / Fastify / PostgreSQL backend using domain-driven design, structured around long-running multi-step business operations. Automated the full market lifecycle — creation, funding, on-chain execution, outcome reporting — turning a manual, error-prone process into an orchestration system that lets operators create 10+ markets a day.

🌐   Community NFT sale — @Atlantis World

Backend services for a sale that raised ~$1M ETH — merkle-tree signature verification, token & NFT airdrop scripts, contracts deployed and verified across Ethereum, Polygon, and Optimism.

📊   Social trading platform — @AltLayer

Frontend lead through several iterations to launch. Hyperliquid integration for long/short perps, Tencent Chat SDK for the social layer, pixelated custom UI as a PWA. E2E coverage including MetaMask wallet signing flows via Playwright + Synpress.

languages      typescript · javascript · go · solidity
frontend       react · next.js · chakra · tailwind · shadcn · react hook form · zod
backend        node.js · nest.js · fastify · hono · postgres · prisma · ddd
web3           hardhat · foundry · ethers · viem · wagmi · privy · gnosis safe · tenderly
ai / agentic   mcp servers · tool & function calling · router orchestration ·
               human-in-the-loop approvals · claude code · codex · custom skills
testing        playwright · synpress · hardhat · foundry · unit · smoke · ci
infra          docker · kubernetes · gcp · aws · vercel · nginx · github actions

I use Claude Code, Codex, and MCP-based workflows every day — not as autocomplete, but as a system. I design task breakdowns, tool boundaries, approval points, and reliability checks so agent-assisted work is something a senior engineer would actually ship. I've built custom Claude Code skill orchestrators with sub-agents (security audit pass over two production products, triaged into an epic, fixes shipped with cross-verification on a local stack).

The differentiator: I don't just use AI tools. I design the orchestration around them — tool surfaces, approval gates, test coverage — so the output is production-ready, not prototype-grade.

I'm actively open to senior full-stack, frontend-heavy, Web3, product engineer, and founding engineer roles — especially with teams building in Web3, AI-native software, fintech, developer tools, marketplaces, or product-led SaaS.

The shape of the work I do best:

• a small team where one engineer can own a meaningful surface area
• ambiguous problems that need a working answer, not a perfect one
• a culture that respects both shipping speed and test coverage

If that sounds like your team — [email protected]. I read everything.