Devpost
Participate in our public hackathons
Devpost for Teams
Access your company's private hackathons
Grow your developer ecosystem and promote your platform
Drive innovation, collaboration, and retention within your organization
By use case
Blog
Insights into hackathon planning and participation
Customer stories
Inspiration from peers and other industry leaders
Planning guides
Best practices for planning online and in-person hackathons
Webinars & events
Upcoming events and on-demand recordings
Help desk
Common questions and support documentation
A great project
Turn your site footage into your training program
Video Intelligence with Spatio-Temporal Augmented Retrieval for Egocentric Understanding
Bringing Spatial Intelligence to the Skeleton of the Building
IronVision Forecaster multi-stage AI system that transforms raw construction site video into structured progress intelligence, next-day task plans, material forecasts, & executive-ready reports.
Reinforcement learning environment to sprinkle spatial reasoning ability onto VLMs.
Splat builds 3D scenes from videos, calculating POV locations & object global positions to build models of scenes accessible through LLM queries, as well as feature highlights and scene exploration.
SiteSense turns raw video into spatial intelligence, modeling activity over time to reveal productivity, idle blockers, and operational insights for real world teams.
Archvision: AI-driven architectural workflows that turn ideas into "Geometry-Aware" 2D concepts and technical artifacts, powering immersive 3D Gaussian Splat environments in minutes.
We turn raw construction footage into structured, supervisor-ready reports that display worker actions, timelines, and productivity metrics by aggregating state-of-the-art AI models.
SiteVault turns messy construction video into precise 3D maps. Using CV2Depth and LLMs, we detect off-spec builds remotely, saving superintendents site visits and fixing AI's spatial blindness.
Fast, realistic room layouts guided by harmony and visualized in 3D. Generate a clean layout from photos, render it with our asset library, and preview flow, spacing, and placement.
Hackathon Project — February 2026 Analyze construction worker body-cam footage with computer vision + AI vision models to power a real-time supervisor dashboard.
Ironsite turns body cam footage into construction site intelligence. Our CV pipeline auto-detects workers, tools, and PPE, reconstructs the scene in 3D, and delivers real-time productivity reports.
An Anchor Calibration Vision Language Model for making better measurements within spatial construction data.
Saving construction managers valuable time and effort with a site management brain
Construction Intelligence using Spatial Evidence and Voice AI
Queryable Memory for Construction Sites
2D camera → 3D spatial sensor YOLOv8 · LLM Semantic Reasoning · No LiDAR
SpatialLens enables true spatial reasoning in VLMs by calibrating images with real-world dimensions and injecting computed measurements for accurate, metric-grounded outputs.
SpatialFlow - Know exactly where every material, worker and zone belongs on your construction site. Every week, in Seconds.
PathGuard is an on-device safety system that detects hazards in a worker’s path and labels site footage with structured safety data, improving real-time protection, reporting, and productivity
We turn ordinary construction video into real progress metrics by comparing geometry over time using full 3D when possible and AI depth when it’s messy. No sensors. No labels. Built for reality.
We propose an ML-powered dashboard so worker shifts become actionable reports. At EOD, the dashboard shows violation logs aligned to OSHA standards, so that safety leads can act, not watch feeds daily
1 – 24 of 28
