Vipul Sharma
AI/ML engineer based in Gurugram. Currently shipping production Angular at Wizni and building agentic systems with Claude — RAG pipelines, MCP servers, and autonomous agents that reason across documents and the web.
I'm an Associate Software Engineer at Wizni Software, where I ship production Angular at enterprise scale.
Outside of work, I build agentic systems with Claude — RAG pipelines, MCP servers, and autonomous agents that combine document context with real-time reasoning.
Long-term, I'm betting on AI engineering as the craft I want to spend the next decade getting really good at.
Building production Angular features at enterprise scale.
Read more Show lessBuilt responsive interfaces for biomedical device dashboards.
Read more Show lessComputer Science — Specialization in AI & ML.
Read more Show lessMultilingual CMS for real-time translation across 60+ languages — engineered as plug-and-play so teams can add multilingual support in minutes. Reduced integration effort by ~95% with a single-line API + NPM SDK; scaled to 10,000+ daily translation requests with 40% faster content publishing.
An autonomous QA system that ingests, normalizes, and reasons over large PDF corpora — combining vector retrieval with multi-hop agents and real-time web search. Processed 120+ PDFs and 1,000+ complex documents at up to 90% contextual accuracy; agents lift answer relevance by 70%.
Natural-language browser automation with zero-dependency export — crawls full apps, builds a persistent semantic memory layer, and emits standalone Playwright scripts ready for CI/CD. Handles Shadow DOM, iframes, and dynamic modals via a ReAct agent + Planner.
Building agentic systems and RAG pipelines in production — MCP servers, multi-hop reasoning agents, and tool-using LLMs deployed across AWS Bedrock and self-hosted stacks.
Shipping web applications end-to-end — React/Next.js on the front, FastAPI and Express on the back, with a focus on real-time features and clean API integration.
Currently learning AWS infrastructure — deploying applications with S3 for storage, EC2 for compute, RDS for relational data, and SNS/SQS for messaging.