Projects Archive

All Projects

A complete view of projects across AI research, product engineering, and deployment-focused systems.

Back to Home
Stock Monk
PythonTensorFlowKeras

Stock Monk

LSTM and NLP-powered stock price predictor that blends market sentiment with historical trends.

Stock Monk – AI-Driven Market Forecasting

A hybrid deep learning pipeline combining LSTM time series forecasting with real-time sentiment analysis from news and social media. Reduced prediction error by 11% compared to vanilla LSTM models.

  • Used VADER and custom sentiment scoring to guide sequence models dynamically.
  • Processed over 50,000 financial headlines and tweets for training sentiment weights.
  • Deployed dashboards visualizing sentiment trends against price movement using Plotly.
SourceDemo
FlutterPythonAWS

Sprout App

Postnatal mental health companion app using AI-driven sentiment monitoring.

Sprout – AI-Powered Postnatal Wellness

A mobile health app that detects early signs of postpartum depression using journal-based sentiment analysis and psychological scoring models.

  • Achieved 40% increase in active users through 15+ interactive features.
  • Implemented NLP models that analyze user inputs to track mental health over time.
  • Improved backend data retrieval by 25% using AWS Lambda and RDS tuning.
  • Targeted for use in rural clinics and NGOs supporting new mothers.
SourceDemo
FlutterPythonScikit-learn

Doctorify App

A disease detection mobile app designed for underserved and rural areas.

Doctorify – Mobile AI for Rural Diagnostics

AI-based diagnostic assistant delivering fast, offline-first medical assessments. Built to serve low-connectivity areas and aid community health workers.

  • ML model trained on 30+ common conditions with 98% accuracy.
  • Reduced time to diagnosis by 70%, increasing early detection rates by 50%.
  • Supports local language instructions and offline caching of results for health workers.
SourceDemo
PythonTransformersLangchain

Sanvad

A multilingual conversational chatbot built using LLMs and Indian language models.

Sanvad – India's Multilingual LLM Chatbot

A context-aware conversational agent supporting 6+ Indian languages, built using transformer models and Langchain.

  • Boosted user engagement by 45% and satisfaction by 50% with culturally tuned responses.
  • Integrated fallback logic, entity memory, and multilingual translation layers.
  • Target use-cases: rural education, digital helpdesks, and language inclusivity in public services.
SourceDemo