GOKUL E

Browser-native metadata removal tool that completely scrubs GPS coordinates, hardware markers, serial tags, and author info from images, video, audio, and PDF documents with 100% client-side execution — zero files ever reach a server.

• Engineered client-side processing pipelines across 4 media formats using HTML5 Canvas, Web Workers, FFmpeg WebAssembly, and pdf-lib.
• Configured lazy-loaded FFmpeg WASM modules with Cross-Origin Isolation headers to support high-speed multi-threaded SharedArrayBuffer memory allocation.
• Integrated comprehensive pre/post EXIF and XMP metadata inspection to verify complete removal before file download.

Full-stack tournament analytics and prediction odds hub for the 2026 World Cup. Features real-time odds integration, comprehensive match schedulers, nation team directories, and persistent user favorites.

• Integrated live Polymarket API streaming real-time market probability odds for upcoming fixtures and championship outcomes.
• Structured complete data repositories for all 48 qualified nations, groups, and tournament phase progressions.
• Implemented smooth interactive UI with responsive standings matrices, fixture countdowns, and quick team search.

Provably fair decentralized gaming application on Arc Testnet featuring 9 smart contract-backed games (Dice, Coin Flip, Plinko, Roulette, Crash, Mines, etc.) using cryptographic randomness.

• Connected Chainlink VRF (Verifiable Random Function) to guarantee tamper-proof on-chain outcome verification.
• Designed session key wallet delegation so players make multiple in-game wagers without repeatedly signing MetaMask popups.
• Supported native USDC stablecoin accounting to eliminate token price volatility during betting sessions.

Trustless peer-to-peer group payment protocol utilizing Solidity smart contract escrow and USDC stablecoin to eliminate settlement disputes and manual payment reconciliations.

• Engineered smart contracts that lock shared pool balances until 100% of participants deposit their allocated split.
• Integrated safety refund mechanisms on group cancellations to guarantee zero fund leakage or rug-pull vulnerabilities.
• Architected Flutter mobile interface communicating with smart contract endpoints through web3dart library.

Convolutional Neural Network image classifier trained on botanical leaf imagery to diagnose diseases and pest patterns across varied crop varieties.

• Constructed end-to-end data pipeline including OpenCV image normalization, data augmentation, and feature extraction.
• Trained CNN architecture with TensorFlow and Keras to classify multi-class pathology categories with high diagnostic precision.