Cloud collaboration for an AR hardware startup

We partnered with the (now sadly defunct) Silicon Valley AR startup Meta View at a pivotal moment in the development of their AR headset. Working directly with their senior technical leadership, we facilitated strategic roadmapping sessions that resulted in a comprehensive 5 year product vision for a cloud collabration platform alongside a detailed 6 - 12 month execution plan. We embedded with their existing engineering team-providing hands-on training, establishing best practices, and architecting the cloud infrastructure on GCP.

The core technical deliverable was a gRPC microservices architecture designed to serve multiple client platforms simultaneously: the Meta AR headset OS (built primarily in Unity 3D), an Android mobile companion app, and web-based tooling. gRPC gave us the performance characteristics needed for real-time AR interactions, but presented challenges for web clients given that browser-native gRPC bindings weren't yet generally available. We evaluated generating GraphQL schemas from our protobuf definitions, but ultimately leveraged GCP Cloud Endpoints transcoding to expose RESTful interfaces to web clients while maintaining gRPC internally. The services ran on a Kubernetes cluster within a VPC, with traffic management handled through ingress definitions rather than a full service mesh, keeping the architecture lean while meeting the immediate scale requirements.

This engagement showcased our ability to operate at both strategic and implementation levels: from multi-year roadmapping with executives to low-level infrastructure decisions around protocol transcoding. We left Meta with a production-ready microservices foundation, a team trained in modern cloud-native patterns, and an architecture designed to evolve as their AR platform scaled.