Skip to content 
The ArangoDB + NVIDIA Hackathon challenged over 1,700 participants from around the globe to build next-gen Agentic Applications using GraphRAG, LangGraph, and GPU-accelerated analytics with NVIDIA cuGraph.
Over four intense weeks, developers reimagined Retrieval-Augmented Generation by fusing graph reasoning, natural language interfaces, and real-world datasets to create adaptive, intelligent systems for everything from urban infrastructure to cybersecurity.
FIRST PRIZE: $15,000 - Qasim Khan
The GDELT Open Intelligence Agent uses LangGraph, ArangoDB, and NVIDIA cuGraph to process natural language queries about global events. Combining structured search, vector and geospatial analysis, and GPU-accelerated graph analytics, it delivers powerful insights through a Streamlit UI - bridging LLM reasoning with real-time, agentic decision-making over massive geopolitical datasets.
SECOND PRIZE: $7,500 - Adam Rahman
AskStreets is an agentic app that answers natural language questions about street networks using OpenStreetMap, ArangoDB, LangGraph, and GPU-accelerated graph analytics. Built with OSMnx and LLMs, it visualizes geospatial queries and routes—empowering planners and researchers to extract insights from urban infrastructure without writing complex code.
THIRD PRIZE : $5,000 - Jones David & Praneeth Yeddu
NeuThera is an AI-driven drug discovery toolkit that integrates generative molecular models, biomedical knowledge graphs, and GPU-accelerated analytics. Built with ArangoDB and GraphRAG, it enables end-to-end compound generation, similarity search, and in-silico validation—empowering researchers to explore drug candidates faster and more efficiently through modular, agentic workflows and scalable graph intelligence
HONORABLE MENTION: $750 - Teddy M
AgentBIM is an intelligent assistant that turns complex building information models (BIM) into interactive, queryable data using ArangoDB and natural language. By converting IFC files into graph structures, it lets users ask plain English questions about spatial elements, relationships, and properties—streamlining access, collaboration, and decision-making in architecture and construction.
HONORABLE MENTION: $750 - Hrishikesh P
DEBy is an AI-powered agentic app that automates data engineering tasks like lineage tracking, ETL scheduling, and impact analysis. Built with ArangoDB, LangGraph, and cuGraph, it leverages enterprise metadata stored in multimodal graphs to help AI agents understand dependencies and perform operations that traditionally require cross-team collaboration.
Public Choice - Vedansh Kumar, Aditya S, Vignesh Karthikeyan, Mohan Ram S
Scopium transforms complex codebases into interactive, graph-based knowledge systems. Using ArangoDB and natural language interfaces, it lets developers query, navigate, and debug code conversationally. By mapping code components and their relationships, Scopium streamlines understanding of large projects—bridging the gap between developers and code through intuitive, AI-powered exploration and insight.
Don’t Miss: Aishwarya Nathani & Nikhil Mankani
ArangoLensAI is an AI-powered crime analysis agent that uses GraphRAG, ArangoDB, and GPU-accelerated graph analytics to answer natural language queries about crime patterns, hotspots, and networks. Built on the Chicago Crime Dataset, it enables real-time insights, relationship analysis, and spatial reasoning to support law enforcement and public safety efforts.