Research Intern @ Samsung R&D | SIH 2025 Winner | AI • Flutter • Backend • DevOps

Agampreet Singh  • Peerlist

Using CNN (Convolutional Neural Network) model, we can predict weather a person has tumor (3 types) or not by submitting a image of CT-Scan or MRI. Model Accuracy - 96.18611931800842 % Model Loss - 23.621788620948792 %

Brain Tumor Classifier

The idea for Censor AI came from the growing demand to make video content safe and appropriate for different age groups without compromising creativity. We noticed that manual censorship is time-consuming and inconsistent. Our goal was to create an AI-driven solution that automates this process efficiently. Censor AI is a smart video censorship tool that: • Automatically analyzes video content using AI. • Applies censorship actions like blurring, trimming, and muting based on selected age categories (6+, 12+, 16+, 18+). • Uses OpenAI to detect inappropriate visuals, actions, or language. • Allows creators and platforms to customize content suitability with minimal manual effort.

Censor AI

Developed a dynamic shopping application and implemented core e-commerce features, cart management and order processing.
Leveraged Google Firebase services for authentication and scalable data management.
Ensured a responsive design and developed a proprietary API for product management.
Integrated the Stripe API for a secure and reliable payment gateway, ensuring smooth and protected transactions.

Crown Shopping

MochaMeetup is a meetup/online chatting system developed using Ruby.
Built entirely on the robust Ruby on Rails framework, ensuring scalability and maintainability.
Utilizes responsive HTML, CSS, and JavaScript on the frontend to deliver a seamless user experience.

MochaMeetup

Built multimodal AI models across a 5-phase NGFW pipeline with 94% DGA detection accuracy; 15μs latency. Leveraged Federated Learning & SOAR automation for 20 real-time threat responses with Zero Trust enforcement. https://sudarshana.tech