Loading…
agentic developer · ai agent architect · senior sdet
I build software through autonomous agents that code, test, deploy, and research around the clock. My job is architecting the skills and tools that make them effective.
stack: claude-code · mcp · playwright · next.js 16 · vercel
01 · PROOF
latest post
How I Built a 3-Agent Autonomous System02 · SYSTEM
Three autonomous Claude-powered agents. They write code, deploy to production, run deep research, and coordinate through MCP servers and custom skills — no human in the loop.
03 · CAPABILITIES
Autonomous multi-agent systems powered by Claude and MCP. Agents that write code, deploy, research, and test — running 24/7 on dedicated hardware.
Tests written in plain English, executed by AI. Natural-language E2E with Shortest, Playwright automation via MCP, and agents that validate their own output.
Software built through agents. Claude Code orchestrates the full lifecycle — design, code, test, and deploy in continuous loops with zero manual steps.
Composable agent capabilities via the SKILL.md standard and MCP servers. Portable, reusable tools that make agents effective across any workflow.
04 · PIPELINE
Autonomous coding agent — writes, reviews, tests, and deploys from a single terminal.
Workflows on every push: lint, type-check, build, deploy. Zero manual intervention.
Bare-metal GitHub Actions runners on dedicated hardware — faster builds, no cold starts.
AI pipeline intelligence — optimizes Actions and catches failures before production.
05 · NOTES