Apoorv Sharma

At Applied Information Sciences, I build Azure-deployed AI platforms that pair Python/FastAPI services with React and TypeScript interfaces. I deliver real-time, role-based workflows across REST and WebSocket endpoints, including secure Azure AD/OAuth2, JWT, and RBAC access controls.

I've built LLM orchestration workflows with function calling and human approval gates, replacing engineering-dependent processes with a self-serve sales-agent experience that generates Process, Proposal, and SOW deliverables in under two minutes. My work also includes real-time voice/AI integrations, multi-user session handling, Docker deployments, and cloud artifact storage.

Previously, I improved a real-time face-swapping pipeline at Gloom Dev through OpenCV-based neck segmentation and skin-tone harmonization, and built a CNN-powered medical image segmentation application during an IBM internship. I enjoy turning AI models into practical, secure applications using Python, Azure OpenAI, FastAPI, React, and Azure infrastructure.