[CVPR 2025] Official PyTorch Implementation of MambaVision: A Hybrid Mamba-Transformer Vision Backbone

Factored inference for discrete-continuous smoothing and mapping.

Official release of MSHT: Multi-stage Hybrid Transformer for the ROSE Image Analysis of Pancreatic Cancer IEEE JBHI https://ieeexplore.ieee.org/document/10006398

The repository gives case studies on short-term traffic flow forecasting strategies within the scope of my master thesis.

A high-performance, extensible aircraft GNC framework for Julia

Monte Carlo simulation for financial instruments.

Using Gaussian Processes for Deep Neural Network Predictive Uncertainty Estimation

COCALITE: A Hybrid Model COmbining CAtch22 and LITE for Time Series Classification

This repository enables training Ultralytics object detection models with the anchor-free LADA assignment algorithm and DFL-based regression. Multi-model feature fusion, joint training, and hybrid inference are also supported.

Source code for the paper "HyPhAICC v1.0: a hybrid physics–AI approach for probability fields advection shown through an application to cloud cover nowcasting".

Production-grade ensemble framework combining XGBoost, PyTorch & Sklearn - 70%+ test coverage with Optuna optimization for time-series prediction

Time Series Forecasting using Linear Regression, Trend, Seasonality, Hybrid models

Final semester research paper analyzing Quantum CNNs vs. ResNet-50 for medical image classification, developed under deadline pressure using AI-assisted coding tools and neural architecture optimizers. Submitted as coursework to demonstrate quantum-classical hybrid vs. traditional CNN efficacy in diagnostic imaging.

Data Science 2023-24

A minimal hybrid Quantum–Graph Neural Network prototype

A machine learning project to develop a robust Book Recommendation System using collaborative filtering and hybrid approaches. The system predicts user ratings and provides personalized book recommendations, evaluated with metrics like RMSE, Precision, Recall, and F1 Score

A robust movie recommendation system using the MovieLens dataset, employing Collaborative Filtering, Matrix Factorization, and Hybrid Models to enhance recommendation accuracy and diversity.

Hybrid quantum–classical neural network for Spanish hate speech detection (HatEval & HaterNet).

In my code portfolio, I generally try new techniques and methods in machine learning. I don't like only copying and pasting.

Forecasted book sales using classical, ML, and DL models (e.g. ARIMA, XGBoost, LSTM)