Givanildo de Sousa Gramacho

I am currently a specialist in automation and control, with a focus on applied research in artificial intelligence and data science. I have over 25 years of experience in electronics, systems development, and the integration of emerging technologies. I hold a postgraduate Lato Sensu degree in Control and Automation from the Federal Institute of São Paulo (IFSP, 2024) and a bachelor's degree in Industrial Automation Technology from Estácio University (2010).My professional background ranges from the development and maintenance of electronic systems to the application of advanced techniques in data science, machine learning, and deep learning. I have completed several specialized courses in artificial intelligence and quantum computing. I have solid knowledge in control engineering, including PID controllers, state-space modeling, and system identification. I am experienced in designing, prototyping, testing, and maintaining embedded systems, developing both hardware and firmware for control and automation applications.I highlight my experience as a research fellow in the PIPE/FAPESP program (TT-4-A) at GAUGIT Soluções, in the field of nuclear engineering, where I developed electronic systems for radioisotope detection and collaborated with researchers at IPEN in creating machine learning algorithms for spectral analysis and classification. I also completed a volunteer data science internship at Data-H, working with probabilistic algorithms and time series for demand forecasting in the supply chain.Currently, I work as a Data Scientist and Researcher at the Center for Artificial Intelligence (C4AI/USP/IBM/FAPESP), serving as a TT-IV fellow (FAPESP #2023/13722-6) in the project Support for Research in Knowledge-Enhanced Learning. I develop fine-tuning and evaluation pipelines for large language models (LLMs), such as LLaMA2 and LLaMA3, both with and without quantization (QLoRA), in addition to RAG (Retrieval-Augmented Generation) architectures. I am also involved in data preprocessing, test set construction, and the application of standardized technical evaluation metrics.I am proficient in tools such as Python, SQL, R, and MATLAB, as well as libraries like TensorFlow and PyTorch, among others, for data analysis, automation, deployment, and monitoring of applications. I have strong skills in mathematical modeling, statistical analysis, and data engineering, with a focus on data-driven systems. I also have experience in system integration using Docker, Kubernetes (K8s), Kafka, and open-source technologies, as well as in cloud environments (AWS, GCP, Azure), performing complete deployments and integrations of AI projects.In addition, I manage projects using methodologies such as Scrum, DevOps, MLOps, LLMOps, CRISP-DM/DS, and Model-Based Design, always aiming for high-quality, timely deliveries aligned with strategic objectives. I am currently conducting research and advanced technical training with LLMs focused on reasoning, deep thinking, and research, integrating these models into autonomous agent applications, with an emphasis on secure deployment and communication through A2A (Agents to Agents) and MCP (Model Context Protocol) standards. (Source: Lattes Curriculum)