DGX Spark Live: NYC Spark Hack Winner feature - A 3D time machine for every building in NYC

NVIDIA Developer · Intermediate ·👁️ Computer Vision ·2w ago
Skills: CV Basics80%
In this livestream, join one of our NYC Spark Hack winners as he shows how DGX Spark powers The Sixth Borough using desktop GPUs to search, reconstruct, and reimagine New York history in real time. DGX Spark acts as the creative “brain server” for the Time Machine, running local LLMs, vision models, and generative tools so viewers can explore archives and neighborhoods as living, navigable stories. What we’ll cover: - The hackathon-winning vision behind Sixth Borough and what made it stand out - How a single builder reconstructed all 1,082,831 NYC buildings using LiDAR and public datasets - Designing a fully local system: running a 30B language model with no cloud or internet - Turning data into narrative: generating building biographies from civic records - Integrating archival media, including NYPL imagery and historical migration flows - Key technical and creative decisions, tradeoffs, and what made it possible under hackathon constraints
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Related AI Lessons

Traffic Light Recognition (TLR) Architecture: 2D Bounding Box Detection
Learn to build a Traffic Light Recognition model using a Fully Convolutional Network and anchor-free approach
Medium · Machine Learning
2D Gaussian Splatting: when removing a dimension makes 3D better
Learn how 2D Gaussian Splatting improves 3D rendering by addressing surface failures
Medium · AI
"Mastering Digital Logic Counters with C++ OOP: A Hands-On Guide”
Learn to implement digital logic counters using C++ and object-oriented programming (OOP) to track events and understand fundamental electronics and computing concepts
Dev.to · Abdullah Fiaz
Como o pensamento computacional me ajudou a estruturar minhas entregas
Learn how computational thinking helped structure deliveries in programming
Medium · Programming
Up next
Yulu Gan - FoundationMotion: Auto-Labeling and Reasoning about Spatial Movement in Videos
Cohere
Watch →