Rupesh Sahani

Developer & Machine Learning Enthusiast

About

I'm a developer and machine learning enthusiast passionate about solving real-world problems through clean and scalable code. I enjoy creating data-driven and full-stack applications — from intelligent systems to tools that make life easier. My core strengths lie in Python, C++, and modern web technologies.

Resume : click here

Skills

C
C++ (STL)
Python
JavaScript
Node.js
React.js
Express.js
Flask
HTML5
CSS3
Bootstrap
MySQL
MongoDB
Git
Linux
LLM's
RAG
LangChain

🚀 Open to Work - Remote & Onsite Opportunities

Coding Profiles

LeetCode
LeetCode
Rupesh528
CodeChef
CodeChef
sahanirupesh528
CodeForces
CodeForces
Rupesh_7
GitHub
GitHub
Rupesh528

Projects

BlogBox

A minimalist full-stack blogging platform inspired by Medium with secure JWT-based authentication and bcrypt password hashing. Uses Prisma ORM with PostgreSQL for robust data modeling and migrations, and a responsive React + TypeScript frontend styled with Tailwind CSS. Backend APIs are built using Express.js with Zod schema validation.

Tech Stack: React • TypeScript • Tailwind CSS • Node.js • Express.js • Prisma ORM • PostgreSQL • JWT • bcrypt • Zod

Online Auction App

A full-stack web application for conducting online auctions with user authentication, real-time bidding, and comprehensive dashboard analytics. Features secure JWT-based login, auction creation and management, bidding history, dark/light theme toggle, and responsive design.

Tech Stack: React 19 • TypeScript • Node.js • Express.js • MongoDB • JWT • Bootstrap 5

ContestBuddy App

A competitive programming reminder app that tracks contests from multiple platforms like Codeforces, CodeChef, and LeetCode. Includes streak maintenance, custom reminders, performance analytics, and motivational streak tracking.

Tech Stack: Flutter • Python • MongoDB • Android Studio

Flight Price Prediction

Machine learning-based web app that predicts flight fares with 85% accuracy. Features a clean Flask UI, data preprocessing (cleaning, encoding, scaling), and performance metrics using MAE, RMSE, and R² scores.

Tech Stack: Python • Flask • Random Forest • Machine Learning

Sentiment Analysis using Transformers

Built a sentiment classification model using pre-trained BERT transformers with 92% accuracy. Fine-tuned models using PyTorch and HuggingFace libraries with an advanced NLP preprocessing pipeline.

Tech Stack: Python • PyTorch • HuggingFace • BERT • NLP

Custom Shell (C++)

A custom Linux shell built using C++ OOP concepts, STL, and templates. Optimized for concurrent process handling, system command execution, and modular scalability for future extensions.

Tech Stack: C++ • STL • OOP • Linux

Contact Search List

A MERN-based contact management app with regex search and real-time updates. Allows adding, editing, and deleting contacts with a responsive, modern interface.

Tech Stack: MongoDB • Express.js • React.js • Node.js