CeLux: High-performance video processing for Python, powered by FFmpeg and PyTorch. Ultra-fast video decoding directly to tensors.

Screw type detection using ESP-EYE, YOLOv5, and TensorFlow Lite Micro for real-time classification on ESP32.

This repository contains a collection of simple yet insightful Data Science projects that showcase my ability to apply analytical thinking, build ML models, and visualize results.

Decision Tree are the typically machine learning algotithm that are work on specific data and generate the tree with yes or no formate.

Open-sourced is_even function powered by LLM

Rank 1st in Public Score with 0.447021 F1-Macro

Real-time cardiovascular risk prediction platform built with Next.js, Flask, and a robust ML pipeline.

Forging a cutting-edge Salary Prediction Software for Software Engineers. Seamlessly blending the art of data science and the precision of machine learning.

REMX API is an automated REM (Random Encounter Model) tool for animal abundance estimation using Machine Learning

This project predicts anime ratings and classifies their success using data from the MyAnimeList API. It applies regression and classification models to explore key influencing factors.

Flight Price Prediction

Official implementation of the TMLR paper: An Analysis of Model Robustness across Concurrent Distribution Shifts.

"Automated data cleaning & validation tool using AI/ML for error detection, correction, and quality scoring."

An interactive Power BI dashboard analyzing car sales data with filters for location, ship mode, and year. Provides insights into sales trends, budget performance, and profitability with key KPIs and visual breakdowns by segment, category, and month.

Machine-learning dashboard to predict whether a cell cluster is benign or malignant using Python and Streamlit.

Privacy-preserving disease risk prediction using the CKKS homomorphic encryption scheme.

A Streamlit-based ML toolkit for data exploration, preprocessing, and classification. Supports custom datasets, model training (CNN, SVM, etc.), and interactive visualization — ideal for rapid prototyping and educational use.

Algorithms and methods focusing on fitting polygonal shapes to the arbitrary point clouds or contours for the spelldraw whiteboard applications.

Customer churn prediction (Logistic Regression). 82% accuracy, recall-focused, business insights included.

ProbNetX is a research-grade Bayesian network framework that learns conditional dependencies from discrete datasets and performs exact inference for predicting missing values.