NAMAN JAIN

At Samsung Electronics, I build and deploy Generative AI systems for real-world engineering workflows, including LLM-powered RAG pipelines, agentic workflows, and production ML models.

I built a scalable LangChain RAG pipeline over tool manuals and waiver data, replacing keyword lookup with semantic search and context-aware Q&A. I also architected an Agentic AI workflow for PPA optimization that automated initial analysis for 80% of cases.

My work includes autonomous log-analysis agents, disk-space and runtime prediction models that improved resource efficiency by 30%, and real-time monitoring for LSF distributed jobs. I work primarily with Python, PyTorch, LangChain, LangGraph, vector databases, Django, and SQL-backed systems.

I previously contributed to critical financial refund and recurring-payment APIs at Juspay Technologies. I’ve also published PEIS 2024 research on graph-based self-supervised learning for molecular olfactory prediction and build efficient models through distillation, compression, and quantization.