AI Street Mapping Portal & On-Demand Capture
Challenge
A global street-level 3-D mapping leader—the Netherlands-based competitor to Google Street View—had just launched next-gen recorders that let customers capture their own AI-ready imagery. To monetise the hardware they needed an online portal where municipalities, utilities and insurers could manage devices, trips and on-demand data processing without phoning support. Legacy tooling shared data only via internal APIs, there was no self-service UI, and every change risked breaking camera firmware or AI pipelines in the field. The solution had to be cloud-agnostic, integrate with existing order-management and image-processing back-ends, and hit a hard eight-month market window
A global street-level 3-D mapping leader—the Netherlands-based competitor to Google Street View—had just launched next-gen recorders that let customers capture their own AI-ready imagery. To monetise the hardware they needed an online portal where municipalities, utilities and insurers could manage devices, trips and on-demand data processing without phoning support. Legacy tooling shared data only via internal APIs, there was no self-service UI, and every change risked breaking camera firmware or AI pipelines in the field. The solution had to be cloud-agnostic, integrate with existing order-management and image-processing back-ends, and hit a hard eight-month market window
Solution
Sigli formed a cross-functional squad (AI + full-stack + mobile) and delivered in three stages.
1. Architecture & discovery – reverse-engineered camera and CV pipelines, mapped REST endpoints, and drafted a micro-service design with clear SLAs.
2. Full-stack delivery – built a React + TypeScript SPA for fleet/recording management, developed Node.js + NestJS APIs, and containerised services with Docker & Kubernetes for blue-green deploys on either Azure or on-prem clusters.
3. On-demand module – a separate back-end app spins up isolated processing stacks (GPU instances, Mapbox tiles, PostgreSQL/PostGIS) per region so clients can request ad-hoc captures.
Dev-ops pipe-line (GitHub Actions → Helm charts) enforces IaC and automated tests; detailed run-books and pair-programming sessions ensured knowledge transfer to the client’s dev team
Sigli formed a cross-functional squad (AI + full-stack + mobile) and delivered in three stages.
1. Architecture & discovery – reverse-engineered camera and CV pipelines, mapped REST endpoints, and drafted a micro-service design with clear SLAs.
2. Full-stack delivery – built a React + TypeScript SPA for fleet/recording management, developed Node.js + NestJS APIs, and containerised services with Docker & Kubernetes for blue-green deploys on either Azure or on-prem clusters.
3. On-demand module – a separate back-end app spins up isolated processing stacks (GPU instances, Mapbox tiles, PostgreSQL/PostGIS) per region so clients can request ad-hoc captures.
Dev-ops pipe-line (GitHub Actions → Helm charts) enforces IaC and automated tests; detailed run-books and pair-programming sessions ensured knowledge transfer to the client’s dev team
Results
Self-service portal live — 120+ B2B users manage 80 + mapping vehicles, plan routes and monitor uploads in real-time.
On-demand provisioning in < 15 min (down from several days) thanks to automated micro-service spin-ups.
Integration complete: order management, image capture and AI-processing back-ends unified via secure APIs; zero critical regressions at launch.
Faster feature velocity: CI/CD pipeline cut release cycle from monthly to weekly sprints.
Scalable & cloud-ready: Kubernetes stack runs on-prem for GDPR but can lift-and-shift to Azure in hours. These wins strengthened the client’s position as the go-to AI street-mapping platform for smart-city and utility projects across the EU.
Self-service portal live — 120+ B2B users manage 80 + mapping vehicles, plan routes and monitor uploads in real-time.
On-demand provisioning in < 15 min (down from several days) thanks to automated micro-service spin-ups.
Integration complete: order management, image capture and AI-processing back-ends unified via secure APIs; zero critical regressions at launch.
Faster feature velocity: CI/CD pipeline cut release cycle from monthly to weekly sprints.
Scalable & cloud-ready: Kubernetes stack runs on-prem for GDPR but can lift-and-shift to Azure in hours. These wins strengthened the client’s position as the go-to AI street-mapping platform for smart-city and utility projects across the EU.