Projects

AI projects built from model to product.

Selected work across real-time multi-modal analysis, computer vision, explainable NLP, plant disease detection, and sign-language recognition.

Computer Vision

AI Gym Rep Counter & Form Analyzer

A real-time fitness tracker that follows body pose, counts repetitions, detects form mistakes, and gives voice feedback while users train.

  • 6 exercisesexercise-specific thresholds
  • Live poseMediaPipe joint tracking
  • 3D guideform reference with voice cues
  • Python
  • MediaPipe
  • NumPy
  • Three.js
  • Web Speech API

Includes phone-as-camera mode, exercise-specific joint-angle thresholds, and a state-machine rep counter.

NLP / Explainable AI

Fake News Detector

An explainable text-classification app that predicts whether an article is fake or real and shows why the model made that decision.

  • LIMElocal model explanations
  • TF-IDFclassic NLP feature pipeline
  • Analysisstyle and vocabulary signals
  • Python
  • Flask
  • Streamlit
  • scikit-learn
  • NLTK
  • LIME
  • TF-IDF

Also includes model comparison and statistical writing-style analysis for sentence structure, vocabulary usage, and formality.

AI + Web App

Plant Disease Detection System

A CNN-based plant-leaf disease classifier packaged behind an inference API for fast image upload, preprocessing, prediction, and confidence scoring.

  • CNNplant-leaf image classification
  • APIbackend inference workflow
  • Scoringconfidence output for predictions
  • Python
  • TensorFlow
  • CNN
  • Flask
  • REST API

Final-Year Project

Mute Motion

A mobile sign-language-to-text system built to make communication easier between sign-language users and people who do not know sign language.

  • Mobile appReact Native experience
  • Vision AIgesture recognition direction
  • Firebaseapp data and backend services
  • React Native
  • Firebase
  • Computer Vision