const jp = {
role: "Tech Lead & DevOps Engineer",
company: "@studiovisual",
location: "Sao Paulo, BR",
focus: ["AI-Powered Automation", "DevOps", "Self-Evolving Systems"],
philosophy: "Best practices only. No workarounds. Ever.",
currentProject: "A Telegram bot that analyzes its own code and creates PRs to improve itself",
};What I actually do day-to-day
- Lead infrastructure and DevOps across multiple clients (AWS, Docker, CI/CD)
- Build AI-powered automation — bots that write code, test it, and deploy themselves
- Architect self-improving systems — the bot proposes improvements, I approve on Telegram, it implements and creates PRs
- Ensure maximum quality — automated testing, linting, coverage thresholds, branch protection
- Manage monitoring stacks — Zabbix, Grafana, Prometheus, custom dashboards
- Optimize cloud costs (FinOps) — AWS right-sizing, reserved instances, cost analysis
Languages
DevOps & Infra
Monitoring & Observability
AI & Automation
Databases & Queues
Testing & Quality
An AI-powered system that analyzes, improves and deploys its own codebase
Cron (daily) Mobile notification
| |
v v
AI analyzes codebase Human approves/rejects
| |
v v
Proposals generated Issue created on GitHub
|
v
Isolated workspace (worktree)
|
v
AI implements the change
|
v
Validates: lint + test + build
|
v
Creates PR --> CI --> Deploy
Stack: TypeScript + Node.js + PostgreSQL + BullMQ + Claude AI + GitHub Actions
Numbers: 400+ tests | 50+ modules | Self-hosted CI/CD | Zero-downtime deploys
"The best code is the one that improves itself."