Oscar Oduor

No description.

A Python-based fraud detection system for checking transactions. Features a training pipeline (fraud.py) using SMOTE, Logistic Regression, Random Forest, and XGBoost for robust model training. The FastAPI backend (app.py) delivers real-time and batch predictions with rule-based explanations.

AgentCodr is a web-based AI coding assistant designed to accelerate the software development lifecycle. It acts as an expert pair programmer and software architect, helping you go from a high-level idea to a fully scaffolded project with initial code in a single step.

A Streamlit-based app for analyzing customer reviews with sentiment prediction and AI-driven insights. The training script (train_sentiment_pipeline.py) preprocesses text data, cleans reviews, and trains a Logistic Regression model with TF-IDF, saved as flipkart_sentiment.joblib.

A public repository for documenting "donations" and handouts by politicians and the total they have done

A compact Python project that predicts short-term stock movement by combining historical price data with news sentiment. Includes a trained model (stock_model.pkl), sample dataset (netflix_stock_data.csv), and a Streamlit demo (streamlit_app.py) for interactive exploration.

MultiStack Converter to convert different stacks to React or Astro

TT

Last Push

Push

Jomi is an Android dating app project (version 5.4) implemented with Android/Gradle and Kotlin. The repo contains an Android app module with UI, resources, and initial Firebase configuration. It’s a working starting point for a modern dating app but remains incomplete.

This project is a Digital Evidence Management System (DEMS) built with PHP and MySQL. It provides a secure platform for law enforcement or organizations to upload, manage, and track digital evidence files.

A compact Python toolkit and dataset scaffold for building object-detection models that identify traffic incidents, obstructions, and vulnerable-road-user crashes. The repo combines dataset organization, annotation utilities, and a YOLOv5-based training/inference pipeline so you can go from raw images to a trained detector quickly.

Personalized movie recommender with Flask API + ML model and React frontend.

pentest_gui.py provides a compact Python-based graphical interface to organize, run, and extend penetration‑testing tasks and utilities. It’s intended as a lightweight, extensible frontend for workflow automation and quick tool access during security assessments.

This project contains a Python implementation of the Deep Deterministic Policy Gradient (DDPG) algorithm and a few example environment scripts.

First Repository