Thomás Freire da Silva Reis's Projects
ETL of foodcenter data using Airflow and PostgreSQL database.
Creating a data environment to load and transform the dataset of ecommerce data. Data extracted from Kaggle.
Experiments with new frameworks using jupyter notebook or not.
Load data from Google Drive, validate the values and format, if it is alright it appends to the final Parquet.
Utilize mlflow to track model parameters and metrics, making easy to monitor and retrain a model if necessary.
✈Pytorch(LSTM) model for prediction of remaining useful life of turbofans of airplanes!
Consult PDF files using LLM model and RAG.
CNN pre-trained model (Resnet18) to extract features of a sequence of images (slices) of CT of Abdominal Region to classify injuries in different organs.
ETL of fact tables of transactions of a sales retail. Later this data was used to build dashboard deployed in Render. Technologies envolved: DuckDB, Plotly, S3 Bucket, Python.
⚖Unbalanced data of diabetes classification
Extração e transformação dos dados de violência contra a mulher da SSP-SP.