Vitor Albres

At iFood, I build the AI engineering behind a major customer-support chatbot handling roughly 3 million monthly post-delivery interactions. I design automation agents, decision engines, and prompt infrastructure that move conversational AI from prototype to reliable production systems.

I've reduced unnecessary human handovers by about 15%, built an order-cancellation decision engine using geolocation, feature-store signals, and business rules, and deployed six internal LLM automation agents. I also helped ship a centralized Prompt Manager for dynamic configuration across customer-support bots and built a unified tickets, chats, and chatbot-data tool now used company-wide.

Before AI engineering, I led resilient authentication migrations and security work at iFood, including a near-zero-downtime AWS migration and MFA design. At Stone, I improved financial-reporting reliability, cut incident response from hours to minutes, reduced reporting incidents, and improved report generation performance by 15%.

I'm deepening my applied AI work through a Master's in Artificial Intelligence at UFPR, researching LLM distillation and fine-tuning. I enjoy solving complex problems where distributed infrastructure, observability, security, cost, and practical AI meet.