Disha Raskar

I am a Computer Engineering graduate and AI/ML researcher, currently working as a Software Engineer at Capgemini, where I focus on building full-stack solutions and exploring cloud security. I have hands-on experience developing hybrid on-device and cloud LLM systems, including a 1.5 GB DeepSeek-R1-Distill model integrated with Llama-rn for offline inference, with cloud fallback for complex queries.

In the Web3/Ethereum space, I have developed ProofDrop, a privacy-focused token drop dApp on Sepolia that allows users to create and claim tokens using secrets without exposing wallet addresses. I also built the DeFi Intent Simulator, an AI-driven Ethereum system where users express high-level intents like “Swap ETH to USDC,” which are converted into execution plans without broadcasting on-chain transactions. Both projects demonstrate my ability to design secure, privacy-conscious, and innovative decentralized applications, integrating React, ethers.js, and Solidity smart contracts.

Additionally, I have worked on EnergyTrade, a blockchain platform for secure energy transactions, and NFT-based projects like Event Ticketing Platform. I have also won in the Unfold 2024 hackathon, Bengaluru, India (Non-Tech Track) for Okto Documentation, showcasing my capability to contribute effectively in interdisciplinary settings. I am passionate about leveraging AI, Web3, and cloud technologies to build scalable, privacy-focused, and user-friendly solutions.