Summary
Mateus Amaral is a senior data scientist based in Rio de Janeiro with 11 years of experience turning business problems into production-ready ML solutions. He blends a business administration background from FGV with deep technical practice in NLP, computer vision, LLMs (GPT series, Llama 2), transformer-based pipelines and vector search/RAG architectures. At Sofya and 1STi he has delivered end-to-end systems—from OCR and visual question answering to intent recognition, entity linking and multimodal embeddings—using tools like LangChain, LlamaIndex and Pinecone. He favors pragmatic, impact-focused models that bridge exploratory analytics and deployable inference, and continuously sharpens skills across speech, text and image modalities. An inquisitive practitioner who describes himself as “trying to understand the world,” he pairs domain-aware thinking with hands-on engineering to extract actionable insights.
10 years of coding experience
Programação de Jogos, Programação, Programação de Jogos, Programação at NAVE / CEJLL
Graduação, Business Administration and Management, General, Graduação, Business Administration and Management, General at Fundação Getulio Vargas
Portuguese, English