Looking for a job

varun kumar

@varunkumar20

Message

AI/ML intern at LightCast building a self-refining knowledge graph for job-description mapping.

India

Message

At LightCast, I engineered a self-refining knowledge graph engine that uses Gemini LLMs and an LLM-as-a-Judge feedback loop to map job descriptions to deduplicated professional taxonomies.

At Infosys SpringBoard, I developed an AI-powered hospital IVR system for symptom triage and ambulance dispatch. I’ve also built a voice-enabled RAG platform and a neural-network digit recognition engine that achieved 92% accuracy without ML libraries.

Experience

Work history, roles, and key accomplishments

AI/ML Intern

LightCast

Sep 2025 - Jul 2026 (10 months)

Engineered a self-refining Knowledge Graph engine using Neo4j Aura and Gemini LLMs, utilizing an LLM-as-a-Judge feedback loop and rejection sampling to extract and map job descriptions to professional taxonomies.

Neo4j Gemini LLM Knowledge Graphs Python Machine Learning

AI Intern

Infosys SpringBoard

Jan 2026 - Apr 2026 (3 months)

Developed a hybrid ILP-AI Hospital IVR system using Gemini 2.5-pro and chirp 3, replacing legacy DTMF with automated symptom triage and optimized ambulance dispatch. Supported 1,000+ concurrent scenarios via a NestJS/Python microservices architecture integrated with HL7/FHIR protocols.

Gemini GCP NestJS Python Microservices HL7 FHIR ILP AI

AI Intern

Ideal Tech Labs

May 2025 - Jun 2025 (1 month)

Architected a scalable full-stack RAG system using FastAPI, Streamlit, MongoDB, Qdrant, and Hugging Face Transformers, implementing voice-enabled semantic search across multi-PDF environments that improved retrieval speeds by 80%.

fastAPI Streamlit MongoDB Qdrant Hugging Face Python RAG REST APIs