Hack the Video Agent Context Graph (OpenAI + Neo4j + AWS + TwelveLabs)
- WHO IT IS FOR: AI, ML, and software engineers interested in building advanced video reasoning agents.
- WHAT IT IS ABOUT: Creating a video agent that utilizes a context graph to analyze and reason over vision, audio, and speech data.
- FORMAT: A full-day technical hackathon featuring expert-led workshops, hands-on building, and live demos.
Unlock the power of video data at the AWS Builder Loft in San Francisco. Video is one of the most challenging data types to search, with critical information often locked inside hours of footage. This intensive hackathon challenges you to build a video agent capable of transforming raw footage into a searchable context graph, allowing for complex reasoning over what is shown, said, and written on screen.
Using a cutting-edge technical stack, you will leverage TwelveLabs for multimodal understanding, Neo4j for graph modeling, and OpenAI for reasoning, all orchestrated by Strands Agents. Whether you are a seasoned AI engineer or a curious builder, you will receive starter templates and expert guidance to help you map entities, scenes, and relationships from raw video into a structured graph.
Join the [SF] HackerSquad for a day of rapid prototyping and innovation. The event kicks off with a series of technical deep dives into the pipeline, followed by a focused hacking session and a final demo window where you can showcase your agent to the judges for a chance to win prizes.