A collection of my favorite works.

Switch

Switch

Switch is a fast-paced, colorful game designed to keep you hooked. You control a bouncing ball, dodging and hitting blocks that constantly move around. Each collision scores points and ramps up the excitement. The gameplay is simple to learn, yet surprisingly challenging to master. It’s the kind of game you pick up for a minute — and end up playing for hours..

Tech Stack:
UnityC#2D Game DevelopmentVisual Studio
BrainyBites

BrainyBites

Brainy Bites is a pixel-style math game that makes learning fun for kids. You guide an adorable baby snake as it slithers toward the correct answers to simple math problems. Every bite helps the snake grow while sharpening young minds. The gameplay is easy to understand, colorful, and rewarding. It’s the perfect mix of learning and play in one charming little game

Tech Stack:
UnityC#Pixel Art Design2D Game DevelopmentVisual Studio
Food-Focus AI

Food-Focus AI

This open-source web app, inspired by HealthifyMe, identifies Indian foods from images and shows their macronutrient breakdown using a custom Vision Transformer (ViT). To improve accuracy, the project uses Mutually Guided Image Filtering (MuGIF) to clean the dataset and remove noise. Features are then extracted using the VGG architecture to analyze food images in detail. A hybrid model combining Vision Transformer and Swin Transformer takes advantage of both methods’ strengths. Hyperparameters are optimized using the Improved Discrete Bat Algorithm (IDBA), making the system faster and more accurate. Tests show the model achieves 99.83% accuracy, far better than existing approaches. This demonstrates how hybrid transformers and smart preprocessing can improve food recognition for nutrition tracking and personalized diet recommendations.

Tech Stack:
Vision Transformer(ViT)Swin TransformerMutually Guided Image Filtering(MuGIF)VGGImproved Discrete Bat Algorithm(IDBA)pythonPytorchTensorflowFlask
Work Ripple AI (Productivity, PWA):

Work Ripple AI (Productivity, PWA):

I built a productivity-focused Progressive Web App (PWA) that combines a Pomodoro timer, to-do list, lofi music playlist, dark mode, and local settings storage for a smooth offline experience. The app is designed for focus and flexibility, with cross-device sync planned for future updates. It’s built using React, TypeScript, Tailwind CSS, and service workers to enable PWA capabilities. The clean UI and responsive design ensure it works seamlessly on desktop and mobile. This setup makes it lightweight, fast, and perfect for everyday productivity.

Tech Stack:
ReactTypeScriptTailwind CSSProgressive Web App (PWA)
PromptPDF

PromptPDF

I built a platform that lets users interact with PDFs using Retrieval-Augmented Generation (RAG). The system embeds PDF content into PineconeDB and uses OpenAI to answer questions in a conversational way. Files are uploaded securely via AWS S3, and user authentication is handled by Clerk. The full-stack app is powered by Next.js 15 with ShadCN UI on the frontend, while DrizzleORM and PostgreSQL manage the backend data. This combination delivers a secure, scalable, and smooth experience for working with PDF documents.

Tech Stack:
Next.js 15ReactShadCN UITypeScriptRAG (Retrieval-Augmented Generation)OpenAI APIPineconeDBAWS S3ClerkDrizzleORMPostgreSQLTailwind CSS
JeevaKrishna's Logo

"I WANT TO PUT A DING IN THE UNIVERSE"- Steve Jobs

Built by JeevaKrishna • 2025