real-time-prediction

Build a production ready Machine Learning service for real-time prediction using FastAPI

Advanced ML system for stock market prediction with real-time data and multiple algorithms

🛰️ Production-ready ML system for geomagnetic storm prediction | 98% AUC, 70% recall | Threshold-optimized ensemble with real-time inference | 29-year dataset (1996-2025) | NOAA SWPC operational standards | Complete MLOps pipeline

Hosting tensorflow keras model using FastAPI. Provide users to upload their own images and call the model to make real time prediction.

This web app uses an AI model trained with TensorFlow.js to guess if a (fake) person would be "focused" or "distracted" based on simulated brain waves (alpha/beta/gamma) generated every second.

This repository contains a project for fruit recognition using the Fruits-360 dataset. The project leverages computer vision techniques, image segmentation, color histogram extraction, and machine learning classifiers to classify fruits into different categories.

This repository contains the coding materials to reproduce the learning error curves of the paper "How Does Data Freshness Affect Real-time Supervised Learning ?".

Ensemble Deep Random Vector Functional Link with Skip Connections (edRVFL-SC) No GPU required • 100× faster training

The project uses computer vision and machine learning to translate sign language gestures into text, MediaPipe detects handprints, and Random Forest Classifier predicts characters in real time. It recognizes all 26 letters of the English alphabet, providing an intuitive interface to improve communication for people who need to use sign language.

Rainfall Detection and Short-Term Prediction using GNSS Signal Attenuation and TinyML on STM32

Heart Disease Prediction App This app uses a trained Random Forest machine learning model to analyze various health parameters such as age, cholesterol, blood pressure, heart rate, diabetes status, lifestyle habits, and more. It predicts whether a person is at risk of heart disease and displays a visual risk score .

Real-time road scene object detection using YOLO11 for autonomous driving perception. Includes training, evaluation, and inference pipelines with a focus on strong accuracy–speed balance and practical deployment.

The Advanced Food Supply Chain Forecasting System is a machine learning-powered application designed to build unified, dual-objective forecasting model for sustainable food supply chains. This simultaneously predicts food demand (orders) and food waste by leveraging real-time external signals such as weather data, fuel prices, economic indicators.

This project centers around two primary aspects: the development of a model for the recognition of American Sign Language (ASL) gestures utilizing a dataset, and the real-time prediction of gestures via a camera feed, accompanied by the conversion of these predictions into spoken language.

Cyclist safety risk assessment application for Austin using Random Forest classification with SMOTE for class balancing. Interactive Streamlit web app predicts cycling incident severity based on time, location, and environmental conditions to promote safer cycling practices.

Built and deployed a Flask-based machine learning system to predict loan default risk using customer demographics and financial indicators. Applied advanced ensemble models like XGBoost and LightGBM to achieve ~99% accuracy. Designed a full-stack solution with real-time prediction capabilities, enabling faster, smarter loan decisions in banking.

American Sign Language recognition using several machine learning models and comparing model performances to Convolutional Neural Network

This repository contains the code for the FastAPI and the Streamlit frontend for the RotorAI project.

Part 4 of the D2C churn project: FastAPI-based churn prediction API that serves the trained model via REST endpoints. Designed for real-time scoring, integration with products, and deployment-ready ML serving.