Giovanni Colonni

Tech — 40% leadership / 60% hands-on full-stack. UI performance overhaul — cut P99 total page load from 17s → 3s (≈83% reduction) by refactoring client-side fetch and data-loading logic, optimising backend API endpoints, and applying targeted React rendering optimisations. Intelligence pipeline (Scrape + RAG + LLMs + Analysis) — end-to-end system for data collection, classification, and clustering using RAG, LLM orchestration, embeddings, and clustering to derive the classification topology. Production pipeline (retrieval, vector storage, eval loop, observability) — powering UX personalisation and product page SEO to cut churn from misaligned messaging.

Feefty is a platform focused on structured financial products, allowing users to buy, request, and create personalised structured products, with additional services for issuers and financial advisors. JavaScript-based stack: React and Next.js on the frontend, Node.js on the backend. Integrated the platform with enterprise financial data providers (Harvest) to keep portfolio positions, quotations, and account status continuously in sync. Designed and implemented a LLM driven system for automated booking. Developed and maintained features across the full product stack — back-office tools and client-facing applications under strict UI/UX constraints.

Designed hybrid CI/CD pipelines for Amadeus applications, from the development phase up to deployment both on Azure and on-premise infrastructure, integrating Terraform in the management of Azure cloud infrastructure and nodes. Developing and maintaining a Container Deployment Manager service for internal tool (both office and production tools) +150 services, leveraging Azure technologies and Services; also Kubernetes resources and other tools like Helm and ArgoCD, React for the FR and python for BK.