NEEMIASBUCÉLI

Master Data Specialist

Curitiba, PR

Built and led the inaugural Data & AI division for a global packaging industry client — delivering forecasting models, ML-driven limits generation, automated tendency detection, and end-to-end MLOps pipelines on Vertex AI, driving significant cost reductions. Supported broader tech deployments via GKE and surfaced model insights through Streamlit and TypeScript applications. For an agricultural fintech client, technically led a data team building a production-ready multi-agent system with Google ADK and Agent Engine, with the user-facing front-end deployed on Cloud Run with SSO authentication.

Senior AI / Machine Learning Engineer

Remote

Led the development of conversational agents across multiple domains. Designed an enterprise support chatbot using LangGraph and Vertex AI — orchestrated as a modular state machine for context-aware multi-turn conversations, deployed on Cloud Run, with semantic retrieval via Vertex AI vector search and integrated with Google Cloud Functions for Google Chat. Architected a voice- and text-enabled automotive AI assistant using LangChain, LangGraph, and LLMs — with modular pipelines on Azure Databricks and MLflow, RAG grounded in company knowledge, voice transcription, text-to-speech, and serverless APIs via Azure Functions.

Data Scientist

São Paulo, SP

Guided organizations including Renault Group in leveraging data science for business growth. Solved forecasting (sales and stock projections) and anomaly detection problems using supervised and unsupervised approaches. Improved material forecasting accuracy to optimize after-sales planning and stock storage. Reduced automotive warranty costs by identifying cost concentrations through anomaly detection. Delivered a full MLOps NLP pipeline to classify homologation sheet impacts using PyTorch and HuggingFace on GCP and Vertex AI.

Data Scientist

Curitiba, PR

Designed and built ML models for a speech analytics product, including Speech Emotion Recognition, Audio Classification, and Voice Activity Detection. Collaborated on exploratory data analysis, feature engineering, and model deployment integrated into the call tracking platform. Worked across NLP and quantitative analytics, using Keras, TensorFlow, MLflow, Docker, BigQuery, and GCP tools.

Scientific Scholarship Holder

Campo Grande, MS

Conducted research on 'Convolutional Neural Networks Applied to Plant Leaf Counting' under Dr. Wesley Nunes Goncalves — work that became a peer-reviewed publication.