Clients — physical security and enterprise operations, APAC
Role — Solution architect and pitch lead
Status — Architecture and pitch delivered across several pursuits · POCs for video analytics and OCR
The scaling problem#
Facial recognition demos work beautifully with four cameras. At 200+ cameras the architecture that got you the demo collapses — bandwidth, inference cost and latency all break at roughly the same point.
The design answer was to stop shipping video to the model and start shipping the model to the video.
Architecture#
- Edge inference nodes — detection and embedding extraction on NVIDIA Jetson class hardware, one node per camera cluster
- Embeddings, not frames — only vectors and metadata cross the network, cutting upstream bandwidth by orders of magnitude
- Central matching service — vector search against the enrolled identity database
- Offline-first operation — nodes continue detecting and buffer results through network partitions
- OTA model deployment — model updates rolled out per-cluster with staged rollback
Related engagements extended the same pattern to video analytics and OCR — container and vehicle identification for a customs tracking proposal, and document extraction pipelines using detection plus OCR at the edge.
Commercial dimension#
I also owned the licensing model — per-camera AI ID licensing versus per-node versus flat site licence — and built the pitch deck and presentation scripts used in client-facing delivery. On-premise deployment was non-negotiable for most prospects in this space, which ruled out several cloud-native vendors and shaped the competitive positioning.