VIPUL SHARMA

• Developed and deployed responsive UI components using NextJS, React Native and Electron improving load time up to 35%
• Integrated APIs with Zustand and Zenoh to enable real-time biomedical data flow, improving device-to-dashboard sync speed by 50%
• Built custom hooks, motion effects (Framer Motion, Reanimated, Skia) and reusable components, increasing frontend code efficiency

• Built full-stack apps with Next.js, Express, Fast API, and LangChain, cutting manual operations by 60%
• Deployed applications on AWS (EC2, S3, CloudFront, SQS, SNS), improving system reliability
• Created lead-generation agents with Neon PostgreSQL handling 10K+ queries per month

Technologies: NextJS, ExpressJS, FastAPI, Langchain, AWS, PostgreSQL, Stripe

• A multilingual CMS available both as API and NPM package for real-time translations across 60+ languages with 97%+ accuracy
• Created plug-and-play translation API and package that reduced integration effort by 95% allowing teams to add multilingual support in minutes
• Scaled backend to handle 10,000+ daily translation requests reducing content publishing time by 40% through optimized caching, routing and workflows

Technologies: Langchain, ChromaDB, TavilySearch, Python3, Docker

• Built an Agentic RAG system that analyzed and retrieved insights from 1000+ complex documents to boost contextual accuracy up to 90%
• Developed autonomous agents capable of multi-hop reasoning, combining document context with real-time web search to improve answer relevance by 70%
• Processed and normalized data from 120+ PDF documents, automating reformatting workflows to improve consistency

• Led backend development for an AI-powered document approval system, streamlining workflows for institutions under AICTE
• Integrated NLP and rule-based checks to validate 100+ document types, reducing manual review time by 70%

• Solved 200+ Data Structures, Databases, Data Science questions on LeetCode and StrataScratch