The problem
Winning education bids requires proving alignment to the exact state standards each district must satisfy — historically a slow, manual task owned by curriculum experts.
What I built
- Architected and shipped a production FastAPI platform (Domain-Driven Design) on AWS — ECS Fargate, RDS, ECR, Secrets Manager, ALB, Cognito OIDC/SSO, and GitHub Actions OIDC CI/CD.
- Built a Claude (Sonnet) tool-use pipeline that extracts standards from PDFs into a 6-level hierarchy; a topological sort plus UUID assignment maps each standard to its parent before insert.
- Enforced bounded-context isolation with import-linter and covered the pipeline with integration, Playwright e2e, and unit tests; a human-review staging queue with content-hash dedup gates every AI extraction before commit.
- Modeled the hierarchy in Postgres and generated stakeholder correlation reports (JSON / Excel / PDF) of per-standard coverage and gaps.
System architecture
🏛️
Architecture diagram — coming soon
FastAPI/DDD · Claude extraction · Postgres hierarchy · AWS ECS/RDS/Cognito · CI/CD. (Upload diagram here.)
Demo
▶
Walkthrough video — coming soon
PDF → extracted hierarchy → product-to-standard coverage report.
Stack
FastAPISQLAlchemy / DDDClaude tool-usePostgresAWS ECS Fargate · RDS · CognitoGitHub ActionsPlaywright
Outcome
Standards alignment, automated at national scale Turned a slow, expert-only task into a repeatable, reviewable workflow.