At Acerto, I architected a multi-agent WhatsApp negotiation system that now drives 2% of total company revenue at near-zero operating cost. I also built the debt-ranking model behind the customer portal, delivering a 26% revenue lift validated through A/B testing.
I've shipped AI systems that resolve 89% of roughly 60,000 monthly support tickets, forecast European beer-market demand at 92% accuracy, and generate multi-million-dollar inventory savings at AB InBev.
At Kumulus, I delivered computer vision, RAG, OCR, lead-scoring, and facial-recognition solutions for enterprise clients across energy, finance, healthcare, and telecom. My work removed hazardous infrastructure fly-overs, automated 80% of receipt processing, and added R$50K+ in new MRR.
Over five years, I've turned models into production systems across fintech, CPG, energy, healthcare, and telecom. I'm an Azure AI Engineer-certified data scientist working in Portuguese and English.
