I design, build and ship agentic systems to production — multi-agent orchestration, RAG, tool use and LLM pipelines on the cloud.
class GiuseppeZileni:
def __init__(self):
self.role = "AI Senior Developer @ Hevolus Innovation"
self.focus = ["Agentic AI", "Multi-Agent Orchestration", "RAG", "Tool-Use / MCP"]
self.languages = ["Python", "TypeScript", "C#", "JavaScript", "Java"]
self.ai_stack = ["LangGraph", "LangChain", "Anthropic Claude",
"Azure AI Foundry", "OpenAI", "Model Context Protocol"]
self.cloud = ["Microsoft Azure", "AWS", "Google Cloud"]
self.arch = ["microservices", "event-driven", "agent workflows"]
def current_mission(self):
return "Turning LLMs into autonomous, reliable, production-ready agents 🚀"
|
Claude Certified Developer — Foundations |
The areas I work on every day with Claude in production:
| Area | What I do | Tools |
|---|---|---|
| 🏗️ Agentic architecture | Agentic loops, multi-agent orchestration, state management and fallbacks | Claude Agent SDK · LangGraph · Microsoft Agent Framework |
| 🔧 Tool design & MCP | MCP servers and tools with clear schemas, permissions and error handling | Model Context Protocol · FastAPI · TypeScript |
| 💻 Claude Code | AI-assisted development workflows: CLAUDE.md, skills, subagents, hooks | Claude Code · Skills · Hooks |
| ✍️ Prompt & context engineering | System prompts, structured output, prompt caching, context control | Messages API · Prompt Caching · Structured Output |
| 📚 RAG & knowledge | Graph-RAG, hybrid retrieval, grounding and citations | Neo4j · Qdrant · PostgreSQL/pgvector |
| ✅ Evals & reliability | Agent testing, response evaluation, guardrails | LLM-as-judge · Test suite · Observability |
Anthropic, Claude and Claude Code logos and trademarks belong to their respective owners. This profile is not affiliated with or endorsed by Anthropic.
Agent Engineering Studio — the lab where I turn ideas into AI agent prototypes ready to test: starter kits, templates, end-to-end demos and best practices to go from concept to first release.
| Project | Stack | Description |
|---|---|---|
| opendata-ai | Anthropic · MCP · FastAPI · Next.js |
Conversational AI over Italian and European open data (CKAN, ISTAT, Eurostat, OECD) |
| knowledge-graph | Graph-RAG · Neo4j · MCP · MAF |
Production-ready Knowledge Graph — companion to the book "AI che Ragiona" |
| mcp-geo-server | MCP · MAF · GeoServer · Claude |
LLM agent that drives GeoServer in natural language (WMS/WFS, PostGIS) |
| limen | Multi-Agent · PostGIS · Python |
Multi-factor landslide risk monitoring across Italy |
➡️ Explore all projects at github.com/agent-engineering-studio
