Summary
Harsh Sharma is a research-driven engineer with 10 years of experience building production-ready AI applications that bridge LLMs, deep learning, and backend integration. Currently at USC and Zage, he designs agentic LLM simulations, policy-guardrail systems, and cloud-deployed FastAPI pipelines while driving data-driven research in the energy sector. Previously at Shell he delivered multiple Azure-based API integrations that produced six-figure operational savings and led GenAI proofs-of-concept to automate IT workflows. Comfortable across Python, PyTorch, LangChain and .NET/C#, Harsh combines ML research depth with pragmatic system design for reliability and scale. He’s a hackathon winner and conference presenter who uniquely focuses on composable agent workflows and audit-friendly ML reliability tooling.
10 years of coding experience