Alessandro Nigro.
I'm a freelance Senior Software Engineer based in Milan, with over 15 years of experience building large-scale software. Today I design and ship AI-powered software to production, on cloud infrastructure built with a DevSecOps approach.
- Golang
- .NET
- Python
- PHP
- TypeScript
- SvelteKit
- React / Next.js
- PostgreSQL
- LLM API
- AI Agents
- RAG
- MCP
- Docker
- Kubernetes
- Terraform
- AWS
- Jenkins
- Vault
What I do today
In recent years I have worked in enterprise settings, with companies in media, manufacturing, banking, telco, automotive, real estate and retail.
AI Visibility & GEO
Consulting to get a brand seen and cited by ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews, plus platforms that measure it: multi-model queries, sentiment, citations and automated reports.
Agents and multi-agent systems
Specialised agents working together on a task, RAG over company documents, MCP servers and benchmarks to measure what actually works. Each model is chosen for the task and its cost.
Generative images, video and audio
Pipelines that combine LLMs, image and music models and automated editing, all the way to publishing. Videos and animations generated entirely from code.
Web intelligence & data
Large-scale scraping and crawling, structured data extraction with LLMs, job queues and dashboards that turn the noise of the web into decisions.
Enterprise architecture
Microservices on Kubernetes, message queues, separate environments up to disaster recovery, and platform migrations inside large groups, with their processes and constraints.
DevSecOps & Cloud
CI/CD with secrets managed in Vault, infrastructure as code with Terraform on AWS, non-root containers, backups and self-hosting.
Vibe coding vs engineering
AI writes code in seconds. Making it run securely, every day, with real users, is a different job.
The generated prototype
- Works in the demo, until real users show up
- API keys and secrets in the code or the frontend
- No tests: any change can break something
- Dependencies and permissions nobody has checked
- LLM and cloud costs nobody planned for
- Code nobody knows how to change anymore
The software in production
- Architecture designed for load, failures and growth
- Secrets in a vault, configuration injected at runtime
- Automated unit, integration and end-to-end tests
- Least privilege, rate limiting, input always validated
- Measured costs and the right model for each task
- Documented code your team can maintain
Built a prototype with an AI coding tool? I'll review it and take it to production.
Write your prompt
Describe what you want to build, the way you would ask an AI. There's an engineer on the other end, though: I'll reply within 24 business hours.