Skip to main content
Krishna Kaushik

Krishna Kaushik

AI Engineer with production experience in RAG systems, multi-agent architectures, and end-to-end ML deployment

Talent
ChennaiMember since July 3, 2026
LLMsRAGOffline EvaluationBenchmarkingPyTorchTransfer LearningImage ClassificationPandasOCIDockerStreamlitPython
·

Hire this person

Send a job offer directly to this candidate

LLMsRAGOffline EvaluationBenchmarkingPyTorch

About

AI Engineer with production experience in RAG systems, multi-agent architectures, and end-to-end ML deployment. Proven ability to architect and ship intelligent systems — from Tri-Modal GraphRAG pipelines to real-time forecasting dashboards — while navigating trade-offs between Generative AI and classical ML approaches. Combines technical depth with the ability to translate complex AI workflows into business-relevant outcomes.

Experience

Machine Learning Intern - XORO Story Technologies Pvt. Ltd.

(2026-01 - 2026-03)

  • Architected and deployed a Tri-Modal RAG pipeline (BM25, vector retrieval, GraphRAG) with a prompt hierarchy including pre/post LLM validation and Graphiti temporal memory; delivered production-grade agentic workflows with accurate, context-aware outputs and reduced hallucinations.
  • Evolved baseline Vanilla RAG (Pinecone) to a Tri-Modal Advanced GraphRAG architecture integrating Neo4j and reranking; implemented Graphiti for temporal graph memory in production-grade agentic workflows.

GenAI Program ML Intern - NeubAItics Research Hub

(2025-01 - 2025-04)

  • Engineered an LLM-powered Semantic Book Recommender using Hugging Face Sentence-Transformers, Zero-Shot Classification, and Sentiment Analysis for multi-dimensional recommendation filtering.
  • Launched an interactive Gradio dashboard for real-time semantic search and visualization of filtered datasets, surfacing data-driven insights for end users.

Education

B.E. - Computer Science and Engineering (AI Specialization) - Sathyabama Institute of Science and Technology (2022 - 2026)

Skills

multi-agent architecturesand end-to-end ML deployment

Reviews

Related