Armaghan R.

Built your app with Claude Code, Cursor or Lovable, and now every fix breaks something else? I take AI-built and AI-powered apps from "works in the demo" to reliable in production. 100% Job Success · 44 jobs · $40K+ earned · 8 years shipping production software · Tech Lead (Backend, Blockchain & AI) at XANA WHAT I FIX • Unstable AI-built MVPs: auth, payments (Stripe), databases (Supabase/Postgres), deployment, and the bugs that keep coming back • AI features that don't work reliably: RAG that gives wrong answers, agents that go off-track, LLM output that breaks your app • Security and readiness before you take paying customers: a written audit of what's broken or risky, then the fixes WHAT I BUILD • AI agents and automation (OpenAI, Claude), RAG over your own data, LLM-powered tools • Production backends in Python, Node.js, TypeScript and Go on AWS and Docker PROOF • As Tech Lead at XANA, I led backend and AI work, including a production RAG system for querying on-chain data • Event-driven microservices, high-volume data pipelines, high-availability infrastructure with failover and monitoring • Web3 infrastructure: full and archive nodes, validators, RPC on Ethereum, Avalanche and Polygon; Solidity and Rust HOW I WORK I start with a short written audit: what's broken, what's risky, what to fix first. Then fixes in small, tested steps. You always know what changed and why. Send me your repo or describe the problem, and I'll tell you what I'd check first.