How Abridge Scaled Developer Infrastructure from Days to Minutes - Blog

Abridge, a rapidly expanding AI company in healthcare, faced significant challenges in scaling its developer infrastructure to accommodate the increasing demand from more than 80 million patient-clinician conversations across 250 complex U.S. health systems. To address this, Abridge restructured its infrastructure using Coder, significantly reducing developer onboarding time from days to minutes by implementing a Terraform-based workspace system on GCP. This system allows multiple developer archetypes, such as clinical scientists and ML researchers, to work efficiently while adhering to different resource and compliance requirements. The architecture is organized into three layers: shared modules, templates, and a control plane managed with Helm, ensuring a consistent and automated environment setup. Compliance and security are embedded into the infrastructure, particularly for workspaces handling sensitive data, by using Istio and strict network policies. The company also tackled GPU resource wastage by incorporating auto-stop policies and right-sizing allocations per template, resulting in reduced costs and increased efficiency. Abridge's approach emphasizes maintaining a limited number of well-defined templates and separating setup logic from template logic to enhance infrastructure stability and compliance, enabling its teams to focus on core product development without infrastructure constraints.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.