Diploma Student in Engineering • Building AI-powered & Offline-first Applications
piyushdas1624@gmail.com • github.com/Piyushdas1624 • linkedin.com/in/piyush-das-7246423a1 • India
Engineering diploma student specializing in leveraging AI APIs (Google Gemini, Model Context Protocol) to build practical, intelligent web software. Experienced in client-side optimization techniques — including DOM virtualization for gigabyte-scale log processing, client-side LaTeX parsing, and local-first browser storage architectures. DEV.to April Fools Developer Challenge Winner.
Languages
Frontend
Backend
AI
Storage
Tools
Challenge Winner • Live Demo • Case Study
Awarded top prize in the official DEV.to Developer Challenge for Volex, a zero-dependency workout volume tracking web app. Features dynamic volume calculation (Weight × Reps × Sets), local IndexedDB data persistence, and is built entirely with Vanilla JS with no external libraries.
Solo Developer • React, TypeScript, Gemini API, IndexedDB, MCP
Personal knowledge management platform with AI organization under active development. Focusing on offline-first architecture, local AI-powered search and tagging, and a clean local-first sync model. Architecture design and core AI integration phases complete.
Solo Developer • Kotlin, Jetpack Compose, Android Accessibility API
Android privacy utility that provides overlay-level screen capture protection using the Android Accessibility API, without requiring root access. Overlay engine and permissions flow implemented.
Full-Stack Engineer • GitHub • Case Study
Built a full-stack AI-powered anime recommendation and personal library tracking application. Orchestrates Gemini Flash for real-time entity parsing and Gemini Pro for narrative context-aware recommendations across 4 external APIs. Built with React, TypeScript, Node.js, and SQLite.
Frontend Engineer • Live Demo • Case Study
Engineered a high-performance browser log parser for gigabyte-scale WhatsApp exports. Combines DOM Virtualization (viewport windowing) with Web Worker background threading to achieve smooth scrolling on chat files that would otherwise crash browser tabs.
Full-Stack Engineer • Live App • Case Study
AI-powered civic complaint reporting platform built for a hackathon. Implements semantic response caching to eliminate redundant Gemini API calls, significantly reducing latency and API costs. Built with React frontend and FastAPI backend.
Software Engineer • GitHub • Case Study
Next.js/TypeScript utility that intercepts AI chat outputs (ChatGPT, Gemini) and compiles LaTeX math equations into native Microsoft Word OpenXML Math (OMML) for fully editable Word documents — no plugins required.
West Bengal, India
Engineering diploma program. Focused on software development, computer science fundamentals, and practical project building alongside academic studies.