Yash Kumar

WORK HISTORY

EDUCATION

ABOUT YASH KUMAR

Most AI demos look impressive until they have to run in production. I focus on building the systems that actually work.I\'m an AI Engineer currently at PayNearby, where I build production-grade LLM and agentic systems. My work sits at the intersection of retrieval architecture, multi-agent orchestration, and real deployment not notebooks, not prototypes.What I\'ve shipped: At GeniOS (Founding AI Engineer), I architected a RAG pipeline using LangGraph and LangChain that improved retrieval relevance by 40% and reduced hallucination across agentic workflows. I integrated Pinecone and FAISS with custom semantic search, cutting query latency by 50%. At HackShastra 2025, my team built CyberSentinel a multi-agent scam detection system for calls, SMS, and WhatsApp in 15 hours. Stack: RAG + LangChain + Whisper + Llama. On-device redaction before any LLM call (privacy by design). Live demo shipped, not a slide deck. I also took 1st place at Qubit Quest, Delhi\'s Esya fest, building with Qiskit on a quantum algorithms problem.Currently I work on agentic system design, RAGAS evaluation pipelines, and moving projects from notebook to production-ready architecture.Open to AI Engineer roles at teams where the work is real and the systems run.GitHub: github.com/Anonymus-Coder2403Email: y••••••••@gmail.com