Projects

Developed a CNN-based plant disease detection model trained on 55,000+ images from the PlantVillage dataset, achieving 91% test accuracy. The system classifies 38 plant diseases, assisting farmers in early detection and prevention of crop infections using AI-powered image recognition.

Developed a Convolutional Neural Network (CNN)-based counterfeit IC detection system that achieved 93.2% accuracy using image augmentation and edge detection techniques. The model efficiently classifies counterfeit and genuine ICs, improving reliability in electronics quality control and security.

Built an LSTM-based air quality forecasting model using 10+ years of real-time EPA data to predict AQI trends in Kansas City. Implemented univariate and multivariate LSTM models, achieving an RMSE of 6.2, enhancing public health awareness and environmental monitoring.

Designed and built a miniature ChatGPT voice assistant using the XIAO ESP32-S3 microcontroller, integrating Google Speech-to-Text and OpenAI’s ChatGPT API. This compact, low-power assistant processes real-time voice queries, achieving under 2-second response time, making hands-free AI interaction accessible on embedded hardware.

Developed Car Guide, an AI-powered search engine that simplifies the car-buying process by allowing users to search, compare, and filter through 5,000+ car models. Utilizing tokenization, indexing, and the Vector Space Model with Cosine Similarity, the platform ensures 89% search accuracy while delivering results in under 0.5 seconds.

Developed a machine learning-powered Health Diagnostic System that achieves 92.15% accuracy in disease prediction. Using Random Forest, KNN, and Logistic Regression, the system provides reliable disease identification based on symptoms, helping healthcare professionals make informed clinical decisions.