• I developed Generative AI solutions for a legal platform, as well as one with similar solutions (an off-the-shelf product). The focus was on chat solutions and a file consultant, using the RAG architecture combined with an intention detector to optimize GPT-3.5-Turbo context problems. I also developed a file summarizer using the FIRAC framework (Facts, Issues, Rules and References, Analysis, Conclusions) and other strategies to handle the GPT context window. Additionally, I integrated the solution with other Azure services, such as document translation and audio synthesis.Link to the news on the Microsoft website -> https://news.microsoft.com/pt-br/machado-meyer-advogados-lanca-ferramenta-com-recursos-de-ia-generativa-da-microsoft/• Proofs of concept for the consistent construction of models aimed at predicting the value of judicial sentences in labor processes, as well as the outcome of these decisions (Python + AutoML + Databricks). • Clustering and ranking of users through unsupervised machine learning and the RFM technique (Recency, Frequency, and Monetary Value) for a company that provides cashback services to consumers, aiming to identify the best users and increase sales for partner companies (PySpark + Databrikcs). • Implementation of projects and proof of concepts (PoCs) using Natural Language Processing to summarize documents and label paragraphs in documents as important or not important for industrial sector companies, streamlining the process of technical manual analysis and budget preparation (Python+ Databricks). • Understanding meeting with clients, hourly budget preparation, presentation of results.

• Project to build text classification models to understand the patterns between process issues and machine breakdown and adjustment cases, focusing on explainability, using the AutoML H2o tool;• Case to predict the risk of a late delivery with “Decision Tree Classifier” in PySpark;• Customer segmentation case based on RFM analysis (Recency, Frequency, and Monetary Value) in R Language;• Case of Extracting Association Rules between customers and products using the “Frequent Pattern Mining” algorithm with PySpark;• Case for customer segmentation with the APRIORI algorithm using R language;• Cases for demand forecasting using algorithms compatible with Spark, and linear optimization algorithms using the PuLP package, in Python;• Credit rating case using Azure Machine Learning Studio;• Participation in the construction of a recommender system using Python, recommending the use of APRIORI and ALS algorithms.

Preparation of procedural documents and legal opinions in the areas of consumer law, children and youth, and crimes in general. In addition to analysis of public civil inquiries and administrative improbity actions.

Assist in the development of custom furniture construction projects, from cutting to assembly and finishing.