Jan 2024 - Present
Teaching Mario to master gameplay through 45,000 episodes of reinforcement learning.
Jun 2021 - Present
Implemented decentralized multi-agent RL algorithms to optimize traffic signal control in large-scale urban networks using neural networks and SUMO simulations.
Sep 2024 - Dec 2024
An AI-driven system using RAG and NLP to deliver personalized product recommendations. Built with Python, Streamlit, and LangChain, it offers dynamic, accurate suggestions via a user-friendly interface and scalable data processing.
Jun 2023 - Oct 2023
Real-time computer vision system with 91.3% mAP accuracy.
Mar 2023 - May 2023
DRL framework for efficient large-scale environment exploration.
Jul 2023 - Dec 2023
Developed a 2D platformer game with unique shooting mechanics, featuring tutorials, power-ups, scoring, and escalating difficulty levels using the Pygame library.
Jan 2023 - Apr 2023
Predicted stock price trends using NLP-based sentiment analysis of news headlines, correlating sentiment with stock movements and training models on historical data.
Jan 2022 - Jan 2023
Engineered a cross-platform meme-sharing app with efficient pagination, optimized Firebase usage, and real-time notifications for 10K+ concurrent users.
Mar 2022 - Aug 2022
Analyzed stock trends of Indian blue-chip companies using moving averages, time series, and density plots to evaluate market behavior and volatility.
Oct 2021 - Dec 2021
Developed an AI-driven solution for maximizing profitability by analyzing price elasticity of demand and forecasting business costs using predictive models on a 20M-row dataset.
Jan 2021 - May 2021
Developed an interactive virtual whiteboard leveraging computer vision to enable users to draw and erase mid-air using hand gestures.
Jan 2021 - Feb 2021
Developed a personalized movie recommendation system by analyzing user reviews from the Kaggle IMDb dataset using machine learning and NLP techniques.
Oct 2020 - Jan 2021
Developed a machine learning mobile app for Charusat University with GPS/NAVIC-based campus navigation.
Jan 2020 - Oct 2020
Developed real-time Facial Expression Recognition software using Keras, achieving 98% accuracy by training a deep learning model on labelled facial image datasets.
Live demo using TensorFlow.js to detect objects through your webcam
Interactive visualization of neural network architecture and training process
Watch AI agents learn in real-time through interactive environments
Exploring techniques for generating secure and efficient code that adheres to modern security standards, including compliance with ISO/IEC 27001 for information security management and ISO/IEC 25010 for software product quality. The generated code is designed to be directly usable for building integrations or SaaS products, ensuring resilience against vulnerabilities and threats while maintaining high-quality outputs.