GSoC 2026 - Reaching out to Mentors of Project #36: Agentic GraphRAG - Prerequisite Task Request

[Abhinav-Avasarala]

Hi Raghu @bharagha and Pankaj @14pankaj ,

My name is Abhinav Avasarala, and I’m a Computer Engineering student at NC State University with a strong focus on AI systems, GPU acceleration, and large-scale data processing.

I’m reaching out to express my interest in Project 36 (Agentic GraphRAG). I’ve been working on agent-based systems and RAG pipelines, including a recent project where I built a multi-agent workflow using LangGraph to process research papers into structured outputs. As part of that, I designed an end-to-end RAG pipeline with document ingestion, chunking, embedding, and retrieval with latency under 500ms.

I’m particularly interested in how GraphRAG systems can be designed not just for correctness, but also for efficiency, scalability, and real-world deployment.

For this project, I’m especially interested in exploring hybrid retrieval strategies that combine vector similarity with graph traversal, as well as evaluating tradeoffs between latency and reasoning quality in multi-hop queries.

I’ve started going through the OpenVINO repositories and would love to get started with a prerequisite task. Could you please guide me toward an appropriate issue or area of the codebase to begin contributing?

Thank you for your time, and I’d really appreciate the opportunity to contribute and learn under your guidance.

Best regards,
Abhinav Avasarala

[bharagha]

We have already several interests on this topic and we have onboarded them to start contributing. I am afraid one more contribution may not be optimal usage of time here. Either you pick another topic that aligns with your interest or you collaborate with those who are already contributing. Let me know what you will prefer.