Web-based Shape & MNIST Recognition Deep Learning Application.

The project is a concoction of research (audio signal processing, keyword spotting, ASR), development (audio data processing, deep neural network training, evaluation) and deployment (building model artifacts, web app development, docker, cloud PaaS) by integrating CI/CD pipelines with automated tests and releases.

CS50 final project - webapp to train neural network

This is a ML based Web App that aims to detect the presence of late blight or early blight on potato leaves, which are the primary causes of crop damage. Additionally, the system recommends appropriate precautions and pesticides to help farmers eliminate the blight and protect their crops and increasing their yields.

API used for image identification and market price retrieval in price finder mobile application

Biodiversity dashboard for reforestation projects - Repo for IBM's Call for Code 2022

audio processing service for mock-buddy

GitHub repository for HackAlliance submission 'CoviFit'

Conditional Generative Adversarial Network (cGAN) model built in Python for generating real 2D images of hands depicting Rock Paper Scissors shapes as used in the game.

The game is afoot, Bitcoin!

A single-page web app serving a Machine Learning model which classifies traffic signs. This Flask app is served by Gunicorn with NGINX as the reverse proxy.

Repository of Classnalytic Machine Learning API. This project for joining NAPROCK2018 Programming Contest, Japan.

Web-app for denoising of text on images

The easiest way to convert images to LaTeX documents!

A SMS spam detection app - Ctrl-Alt-Debt Hackathon

アップロードしたフォントと似ているフォントを教えてくれるアプリケーション

This project is a concoction of research (audio signal processing, keyword spotting, ASR), development (audio data processing, deep neural network training, evaluation) and deployment (building model artifacts, web app development, docker) by integrating CI/CD pipelines with automated releases and tests.

This application is designed for environmental audio classification and aims to assist individuals who are deaf or have severe hearing loss by focusing on environmental and urban sounds.

Spam classifier using LSTM-Recurrent Neural Networks. Deployed using Tensorflow Serving and Flask, built with Docker Compose.

skin-lesion-ML is a mobile app that takes photos of users skin lesions and sends it through a Machine Learning pipeline to discover cancer risks.