Neo4j Graph Data Science Specialist and AI Architect with 15+ years of experience in ML, NLP, LLMs, and agency architectures. Creator of the Graph Mind and Code Idea methodology, integrating LLMs, Retrieval-Augmented Generation (RAG) and grounding on Neo4j graphs to provide contextual, explainable and scalable AI solutions.

Specialist in fine-tuning and production deployment of AI models (preferably AWS, Azure, GCP), with a proven track record in experimentation, evaluation and optimization. Highly skilled in Python, SQL/Cypher, and Graph Data Science (Neo4j GDS), applying a data-driven mindset to deliver measurable impact in critical domains such as finance (Spend/Compliance) and healthcare (diagnostic support).

Experience in technical leadership, interdisciplinary collaboration, and delivering high-value products aligned with corporate requirements for performance, reliability, safety, and compliance.

Key competencies

AI/ML: Machine Learning, Deep Learning, NLP, LLMs, Agentic Systems, Fine-tuning, Distillation, RAG, Grounding Techniques, Explainable AI, Model Optimization Programming: Python, FastAPI, Flask, SQL, Cypher, PyTorch, TensorFlow, Hugging Face, LangChain, scikit-learn, JavaScript, TypeScript, Next.js Data & Graph: Neo4j, Graph Data Science, Graph Databases, Knowledge Graphs, Embeddings, TF-IDF, Cypher Queries, Link Prediction Nuvem e MLOps: AWS (SageMaker, Lambda, ECS, ElastiCache), Azure, GCP, Docker, Kubernetes, CI/CD, Vercel, Implantação na nuvem Visualização: Cytoscape.js, D3.js, painéis interativos, visualização de gráficos Domínios: Concur Spend AI, Finanças, Gerenciamento de gastos, Conformidade, Detecção de anomalias, Arquitetura multilocatário

Professional experience

Scientist / Founder of Machine Learning – Graph Mind 03/2013 – Present

•  Designed and implemented a knowledge synthesis platform integrating LLMs, RAG, and Graph Data Science to provide contextual, auditable, and actionable answers across the finance, compliance, and healthcare domains.

• It built a multimodal semantic graph with 45,000 nodes and 400,000 enriched relationships (TF-IDF scores, PubMed occurrences) with all 10,000 diseases and more than 5,000 symptoms.

• Applied medical ontologies and bibliographic data to increase accuracy and explainability.

• It has established partnerships with industry programs: NVIDIA Inception Program, Neo4j Startup Program (AuraDB Enterprise), Microsoft for Startups Founders Hub (early access to RAG).

Data Scientist / Systems Architect – TOTVS Health, Porto Alegre – Brazil 03/2011 – 03/2013

• He architected the development of a diagnostic support system for the HCor Hospital, São Paulo.

• Led the implementation of electronic patient record systems, ensuring performance, reliability and compliance with health standards.

Education

• PhD (incomplete) – Medicine/Neurology – AI Focus – PUCRS – 2016 Thesis: Computational model to support the diagnosis of epilepsy. Advisor: André Palmini.

• M.Sc. – Computer Science – PUCRS – 2007 Thesis: Indexing architecture applied to PACS servers for medical image processing. Advisor: Eduardo Augusto Bezerra.

• B.Sc. – Computer Science – PUCRS – 2004 Thesis: Storage model for retrieval of medical images with diagnostic support.

Academic and professional contributions

• Revisor: Journal of Health Informatics (JHI), Journal of Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization (JCM-BBE: IV).

• Member of the Technical Programming Committee: Brazilian Symposium on Applied Health Computing (SBCAS), Brazilian Congress of Health Informatics.

• Moderator: Technical Session on COVID-19 – ERCAS/USP.

Professor of Python and Artificial Intelligence – University Technology (03/2013 – 03/2023)

• He has designed and taught undergraduate courses in Algorithms and Programming, AI (Python), Computational Logic, Object-Oriented Programming (Java), Database, Systems Analysis and Project Management.

• He taught postgraduate courses (2019–2023) in Artificial Intelligence, Deep Learning, Machine Learning and Business Intelligence.

Certifications

• Neo4j Graph Data Science Certification

• Building Knowledge Graphs with LLMs using Neo4j and LangChain

• Building Neo4j applications with Python, Java, and NodeJS

Languages

• Portuguese: Native

• English: Advanced

Porto Alegre, RS – Brasil | +55 51 99113-1972 | dan@graph.mdLinkedIn: linkedin.com/in/graphmd | GitHub: github.com/graph-md | Portfolio: graph.md