Rajarshi Paul

An android mobile application designed to track fitness journey over the period of time. It Features Real-time calorie count bmi and body fat % calculation . Provides workout plans with flexibility to let the user build their own workout plans with calories burnt per work session gets added up and refreshed every week for better tracking . Fitness data of each body part gets vizualized with curves and charts . Integrated an AI Assistant to give workout insights and suggest next planning or weightlifts . 1)Backend hosted on a cloud server 2)Testing done with expo go 3)APK built using expo packages and android studio .

Developed Nexus, a web-based AI-powered productivity and learning platform designed to streamline professional and academic workflows. Nexus integrates multiple tools into a single interface, enabling users to save time, enhance learning, and create polished outputs efficiently with a custom domain . We hit a milestone of 1000+ users recently Key Features: Automated Reports & Charts: Upload raw data to get it cleaned and then transformed into dashboards and exportable PDFs. AI Mentor Mode: Converts PDFs into Q&A, key points, and quiz content to facilitate rapid learning. Interview & Quiz Mode: Helps users prepare for interviews or self-assessment using AI-generated questions. CV Builder: Generate professional, ready-to-use CVs with AI assistance. Web Builder: Create functional web pages directly from the platform.

A full-stack system that predicts stock prices and shows interactive charts for accurate data tracking and provides external market info using gemini api , built using FastAPI + XGBoost with technical indicators , moving averages, and volatility. It: • Fetches live market data and predicted data with interactive charts • Trains an ML model dynamically • Forecasts future price ranges • Generates confidence intervals • Classifies trend sentiment Also it contains a gemini assisted chatbot that provides market and other extrernal information as well . The chatbot has some features like web search and image fetch using respective APIs Deployed it to production and handled real-world issues like dependency conflicts and cloud runtime debugging. Working on improving forecast realism using recursive feature updates next.

Developed and deployed a mobile-first AI web application designed to eliminate food waste by generating instant recipes based on available ingredients. The app leverages computer vision to provide a "point-and-cook" experience for users. Key Features: 1)Zero-Friction UX: Designed without a login system to allow users to get recipe results in seconds. 2)Gemini Vision Integration: Utilizes Google Gemini 2.5 Flash to analyze photos of refrigerators or pantries and 3)identify ingredients with high accuracy. 4)Comprehensive Recipe Output: Generates five distinct recipes including step-by-step instructions, caloric content, and full nutritional breakdowns. 5)Mobile-Optimized Interface: Built with a "mobile-first" philosophy to ensure a seamless experience on smartphone browsers.

Mobile-first recipe generator built with React, Google Gemini 2.5 Flash API, and Vercel. Uses Gemini Vision to detect ingredients from photos and instantly generates five recipes with nutrition breakdowns and step-by-step instructions.