WorkMcGraw HillDraft, not yet approved
A federated GraphQL backend and AI search for McGraw Hill's Sharpen app
We helped stand up the backend for Sharpen, McGraw Hill's study app: a federated GraphQL service, the deployment and test infrastructure behind it, and OpenAI search over a vector database.
- Backend we helped build
- Sharpen study app
- Deploy and integration test pipeline
- Docker, Terraform, GitHub Actions
Challenge
McGraw Hill was launching Sharpen, a new study app, and needed its backend built. The service was a federated GraphQL backend, and the hard part was operational. Deployments and pipelines had to be streamlined, and integration tests had to run efficiently across the service.
What we built
We helped hire and onboard the backend team, and then built with them.
Our part covered the federated GraphQL service, new microservices, and the infrastructure that deploys the backend and runs its integration tests. We also hooked OpenAI up to a vector database to improve search.
How
The services are Node.js, Express.js and TypeScript, backed by PostgreSQL on AWS RDS and Redis, with Kubernetes and New Relic in the stack.
We set up deployments and integration test runs with Docker, Terraform and GitHub Actions.
Results
Sharpen got a solid backend codebase and infrastructure, and a well-optimised API. The OpenAI and vector database work laid the foundation for better search in the app.