AI Developer & Founder · Automation Builder
I build chatbots, automations and dashboards that cut operational time and give back data control. Real outputs, not demos. Measurable from day one.
I’m an Italian AI Developer and Founder. I build practical tools that combine artificial intelligence, automations and full-stack development. I work on systems that cut repetitive tasks, give back data control and accelerate operational decisions.
I’m enrolling in a Computer Science degree to consolidate the theory behind what I already build in production. My background is technical-aeronautical — precision, complex systems, attention to detail are things I bring to the code.
I work in Italian (native) and technical English (AI/API documentation). Italy, remote-first, open to concrete collaborations.
Claude Code skill that delivers code review like an Italian grandmother: no sycophancy, push-back on red-flag requests, long-term code health awareness. 3 modes (Full / Lite / Del Sud), 4 slash commands, public reproducible evals. 50% sycophancy reduction, 12.5× more push-back vs default Claude on risky prompts.
Chrome extension that auto-fills any web form from your personal data. Multi-document, anti-hallucination guards, semantic mapping via OpenAI with strict scrubbing on sensitive fields (IBAN, passwords, CVV). Vanilla TypeScript + Vite + Zod, no UI framework. 6-phase architecture with AI Pass 1+2 and anti-hallucination validation.
AI assistant integrated for client requests and corporate document lookup, with automated data handling. Reduced repetitive requests, faster access to information.
Interactive dashboard with automated services and dynamic content. Platform with AI assistant and external API integration for simple content management.
Login area, client dashboard and chatbot for automatic updates. Centralizes clients, work progress and communications — fewer manual updates, cleaner process.
Project status monitoring with secure access, notifications and intuitive interface. Cuts ping-pong communications, continuous monitoring.
Contextual assistants with document retrieval (RAG) and natural responses on real corporate data. They handle frequent requests autonomously, integrating with existing systems.
Workflows between email, CRM, databases and APIs to eliminate manual steps and reduce process errors. Reliable orchestration with n8n and custom code where needed.
Platforms with login, user management and real-time KPIs. Solid backend API + reactive frontend to monitor activity and performance without fragility.
Initial discovery with priorities, timelines and concrete technical roadmap. For those who want to start cleanly without burning budget. Output: document + working prototype.
Automation isn’t magic, it’s engineering.
I cut the superfluous to leave room for value. Every component exists because it removes real friction, not because it’s nice to have.
Example: in the RDD dashboard, the chatbot replaced 4 manual touchpoints with 1 structured conversation, reducing internal response times.
I don’t use AI because it’s trendy.
I use it when it solves problems that were previously impossible or too expensive. When a heuristic does the same job, I write the heuristic.
Example: in Universal Form Compiler, before calling the AI I run a local fast-path (autocomplete, name=email). Only remaining fields go to the model, and anti-hallucination is system-level guard, not optional.
Every project starts from a KPI.
Response time, hours saved, errors reduced, conversions up. Reports are automated and readable by management too.
Example: for every chatbot I track auto-resolution rate, average latency, accuracy on real test sets. Numbers land in a dashboard weekly, no need to ask.
Medium-complexity chatbots or n8n flows take 2-4 weeks: discovery, prototype, real-data testing, release. Full web dashboards: 4-8 weeks.
No. We map data, sources and APIs to integrate together. If you have CRM or help desk I plug in, otherwise I propose a lightweight stack (Supabase + FastAPI) from scratch.
I set KPIs upfront: response times, cost per ticket, conversions, team hours saved. Automated weekly reports readable by non-technical stakeholders.
Remote-first from Italy. Available for on-site discovery or kick-off sessions, async work for the rest. Standard tools: Slack/Teams, Linear/Notion, Loom for demos.
Python for AI/automation backend, TypeScript for frontend and tooling, FastAPI + Supabase as a fast stack, n8n for workflow orchestration, OpenAI/Anthropic for LLMs. Pragmatic choices, optimized for time-to-release.
Yes. Nonna (Claude Code skill) and Universal Form Compiler (Chrome extension) are open source on GitHub under MIT. For NDA enterprise projects I share architecture summaries and demos.
Tell me what you want to improve. I’ll reply with a concrete proposal — goals, stack, timeline, next steps.
raffaelefrancesco.damato@gmail.com“Real outputs, not demos. Measurable from day one.”