query-decomposition

Agentic RAG system for complex document intelligence using multi-source retrieval across vector search, GraphRAG, and SQL. Includes query transformation, citation-grounded answers, self-reflection, corrective repair, RAGAS-style evaluation, conversational follow-ups, and a deployed frontend on AWS.

GraTAG — Production AI Search via Graph-Based Query Decomposition and Triplet-Aligned Generation with Rich Multimodal Representations

Drowning in documents? Stop searching and Start Chatting! Ask questions, get follow-ups, and generate summaries across multi-format documents using an advanced RAG system with OCR.

Query optimization case study demonstrating execution-plan analysis, query decomposition, materialization, and staged data processing to improve optimizer accuracy and performance.

RAG system implementing query expansion, retrieval enhancement, context compression, and comprehensive LLM-as-a-judge evaluations.

A fine-tuned Qwen2.5-1.5B model that breaks complex multi-hop questions into atomic sub-queries with dependency graphs, doubling RAG retrieval accuracy in ~50ms