Caixabank Data AI Hackathon submission Permalink
Data Wrangling, XGBoost fraud detection model, LSTM expense forecasting with PyTorch, and an AI agent using Langchain and a locally hosted llama model for PD...
Posts by Year
2024
From Local to Cloud - Mastering ML Model Deployment
Explore the journey of deploying machine learning models from local environments to cloud platforms.
Multipeak Wavelength Detection of Spectrally Overlapped Fiber Bragg Grating Sensors Through a CNN-Based Autoen-coder Permalink
Journal paper presenting an unsupervised approach for obtaining the spectral positions of spectrally-overlapped fiber Bragg grattings with experimental valid...
2023
RSNA Abdominal Trauma Detection AI Challenge entry Permalink
Pytorch models for RSNA ATD AI Challenge, together with preprocessing using Pytorch and Monai for data loading and augmentations of volumentric data.
Building a Personal Journal App with MongoDB and Streamlit
In today’s digital age, maintaining a personal journal has evolved from the traditional pen-and-paper approach to digital solutions that offer convenience an...
Small business data intake and dashboard Permalink
A webapp built with streamlit, with cockroachDB (postgreSQL).
Peak detection of spectrally-overlapped fibre Bragg gratings using an autoencoder convolutional neural network Permalink
Conference paper presenting an unsupervised approach for obtaining the spectral positions of spectrally-overlapped fiber Bragg grattings (a type of fiber opt...
2022
Peltier Temperature Controller Permalink
A PID arduino controller for a peltier for temperature precise control.
2020
New feature for metric-learning package: Sparse Compositional Metric Learning Permalink
Accepted pull request for a major feature proposal for the python package metric-learn.
Urban Audio Classification Permalink
This work was part of the final project for IPD423 - Advanced Signal Processing. The following report was the output from this endevour and summarizes the ch...
Stock Market Forecasting Permalink
Stock market prediction using a GAN with an LSTM as generator and a FNN as discriminator. At inference the generator is used as a predictor.