Back to case studies
Photogrammetry + Cesium + cloud GPU
WaveForm 3D Reconstruction Platform
A photogrammetry and Cesium 3D platform for time-indexable reconstruction, asset inspection, and preventative maintenance workflows.
My role
Lead full-stack engineer responsible for capture architecture, reconstruction orchestration, Cesium integration, cloud delivery, GPU processing, and observability.
Overview
WaveForm connects commodity capture, cloud reconstruction, and browser-based 3D review into one workflow for equipment condition tracking.
Problem
Industrial inspection teams need time-indexable 3D reconstructions they can revisit, compare, and use to target preventative maintenance before equipment failure.
Architecture
- Mobile capture flows collect large inspection media from commodity hardware.
- FastAPI services coordinate upload metadata, reconstruction handoff, and model status.
- AWS Lambda/GPU processing handles heavy photogrammetry reconstruction work.
- Cesium renders delivered models in a browser-based 3D geospatial viewer.
- Sentry monitoring makes failures visible across capture, processing, and viewer surfaces.
Key features
- Photogrammetry reconstruction
- Time-indexable 3D outputs
- Cesium 3D model viewing
- AWS Lambda/GPU processing pipeline
- Cloud model delivery
- Production error monitoring
Screenshots
Product surfaces and system views
PhotogrammetryCesiumAWS LambdaGPU ProcessingFastAPIReact NativeSentry
View 1
View 2
Results
- Technically deep capture to reconstruction to Cesium viewer workflow.
- 3D outputs aligned to asset inspection and preventative maintenance use cases.
- Cloud processing path for heavyweight model generation and delivery.
What I improved
- Turned captured media into browser-accessible 3D inspection assets.
- Separated capture, reconstruction, and viewer concerns.
- Made the architecture observable enough for production support.