John Varshan J

AI Engineer | Generative AI, RAG & Agentic AI | Vector Databases & Production AI Systems | Ex-I-PAC

Bengaluru, Karnataka, India

About

I am an AI Engineer passionate about building production-grade AI systems that solve real-world business challenges. My work focuses on Generative AI, Retrieval-Augmented Generation (RAG), Agentic AI, intelligent search systems, and scalable AI applications. At I-PAC, I developed a large-scale RAG platform processing over 3.7 million IT incident records using PySpark ETL, semantic chunking, vector embeddings, and PostgreSQL pgvector. I also built an intelligent multi-model LLM routing framework that reduced inference costs by 64% while enabling real-time incident analysis and decision support. In addition, I have worked on AI-powered automation solutions, document intelligence systems, semantic search applications, and workflow orchestration using modern AI frameworks and tools. My experience spans the complete AI lifecycle, from data processing and retrieval to model integration and deployment. My technical expertise includes Python, SQL, Machine Learning, Large Language Models (LLMs), RAG, LangChain, LangGraph, Agentic AI, Vector Databases, Prompt Engineering, LoRA/QLoRA Fine-Tuning, PySpark, Streamlit, and AI System Design. I hold a Master's degree in Applied Data Science from SRM Institute of Science and Technology, where I built a strong foundation in statistics, machine learning, data engineering, and applied AI. I am passionate about transforming cutting-edge AI research into scalable products that create measurable business value and improve decision-making. 📩 Open to opportunities, collaborations, and discussions around Generative AI, AI Engineering, Agentic Systems, and Machine Learning.

Experience

  • Associate AI Engineer at I-PAC (Indian Political Action Committee)
    Jan 2026 - May 2026 · 5 mos

  • AI Automation Engineer Intern at Neural Transformers.ai
    Aug 2025 - Sep 2025 · 2 mos

  • Data Analyst Intern at Indian Meteorological Department
    May 2024 - Jun 2024 · 2 mos