I spent years as a PHP developer. Then I realized the language was never the point.
The point was knowing what to build, how to structure it, and why it should work that way. AI just removed the part where I had to remember semicolons.
- AI agents & automation. From "we do this by hand every day" to a deployed agent that does it: multi-agent workflows, MCP servers, LLM integrations wired into real business processes
- Scraping & data pipelines. Parsers and ETL that keep working in the messy real world; clean delivery as APIs, exports or dashboards
- Backend & integrations. APIs, third-party integrations, legacy rescue. Laravel, Go, Python
AI Agents MCP Claude Code Python Go Laravel Docker PostgreSQL Redis Ansible
Systems thinking. Knowing why your database will choke at 10k RPS. Understanding that "just add a cache" is not architecture. Years of debugging production at 3 AM taught me how systems actually work.
I design systems and build them with AI agents: architect the solution, orchestrate the implementation. Small fixed scopes, short feedback loops, working software over promises. The dev workflows I've built make the whole process feel unreasonably fast.
Turns out, the best use of a decade of engineering experience is telling AI exactly what to do.
🟢 Open to projects · 📫 sn.one.dev@gmail.com