Commissioned Engineering for a Startup: From 3D Model to Automated Quantity Takeoff
Why startups and SMEs turn to our team for complex custom software: An engineering case report from our commissioned client work on QTO-AI.
The Client Startup's Challenge
When Standard Web Agencies Hit a Wall with Complex Domain Formats
Many founders and mid-sized companies have a clear domain vision, but face a steep technical barrier: Their data lives inside specialized formats (3D CAD files, industrial telemetry, or legacy SQL databases) that standard web agencies cannot parse.
Commissioned as custom software development for a specialized startup (QTO-AI), our task was to engineer an automated parser and calculation worker pipeline that ingests 3D building models, extracts exact geometric quantities, and outputs structured tender documents.
Python Workers + Graph Database + AI Classification
Acting as the client's full-stack engineering partner, we built a decoupled microservice and background worker architecture — following the same "New Digital Yarn" philosophy we use to weave disparate data sources into Astro BI dashboards:
- Dedicated Geometry Parser Worker: An asynchronous Python worker extracts wall areas, slabs, openings, and insulation thicknesses directly from uploaded 3D models.
- Memgraph Dependency Graph: Spatial relationships between components are mapped inside a graph database to derive required secondary line items.
- AI Classification & Export: A multi-stage rule and AI engine maps quantities across 20 trades and exports standardized exchange files (GAEB X83/P83 & PDF).
Explore the full technical breakdown in the QTO-AI commissioned engineering case study.