Shivam Deepu Prajapati
Full Stack Engineer
Experience
Solana SuperDev Fellow
Solana Foundation
Remote
Chosen as 1 of 120 fellows for the Solana Super Dev Fellowship funded by Solana Foundation. Built a Tip Link-inspired SOL payment platform using Rust and Anchor with wallet-free transfers.
Designed a 3-node key management system using Shamir Secret Sharing for distributed key storage and recovery. Implemented key rotation and distributed secret management to enhance wallet security and fault tolerance.
Full Stack Developer Intern
Uknowva HRMS
Mumbai, India
Redesigned the Graddi. ai platform, enhancing UI, UX, responsiveness, accessibility, and scalability. Reduced CLS from 0.35 to 0.04 and LCP from 4.8s to 2.1s through targeted Core Web Vitals optimizations.
Migrated the frontend from Daisy UI to Shadcn UI, increasing consistency, component reuse, and maintainability. Tested and calibrated OCR pipelines using Open Router, boosting document extraction across diverse inputs.
Education
University of Mumbai
Bachelor of Engineering (B.E.)
Current CGPA: 8.99
Skills
C++TypeScriptRustReact.jsNext.jsNode.jsExpress.jsSocket.ioTailwind CSSShadcn UIAnchorGitGitHubVisual Studio CodePostmanRemixSolana Web3RedisSQLPostgreMongoDB
Projects
Oryxen
Anchor, Next.js, Rust, Solana
Built a Solana perpetual aggregator to route orders for derivatives on-chain protocols Drift and GMXSol. Used Drift and GMXSol SDKs and IDLs to facilitate order routing and execute on-chain transactions. Created a composite trade frontend with real-time data, enjoyable UX and rapid access to market liquidity.
Pump.Fun Indexer
Rust, Redis, PostgreSQL
Built a real-time Pump.fun trade indexer streaming Solana blockchain data via Helius WebSocket API. Extracted token metadata, bonding curve progress, holder balances, and buy/sell trades with USD pricing. Designed a two-process Rust architecture (ingester + worker) using Redis pub/sub for scalable indexing.
Merkle Subscription System
Rust, Anchor, Solana, PostgreSQL
Built a scalable subscription verification system using SHA256 Merkle trees with O(log n) membership proofs. Reduced on-chain storage by 99.9% by storing only a 32-byte Merkle root instead of individual subscriber data.
Intelli-Mind
Next.js, CodeMirror, Lexical
Built an AI-powered coding interview practice platform with real-time problem solving and instant feedback. Developed using Next.js and CodeMirror to create a dynamic browser-based coding environment.