
Document Anonymization Platform.
An asynchronous pipeline for anonymizing sensitive documents, built to be trusted from upload to download
Year.
2025
Employer.
Sagacify
Role.
Tech Lead
Context.
A platform for anonymizing sensitive data before it can be processed or shared: upload a document, choose which data types to anonymize, run the anonymization, then review the result side by side with the original before downloading it.
Challenge.
Ship a reliable, production-ready anonymization pipeline: a backend that handles large documents asynchronously, and a review step trustworthy enough that people actually rely on it. Own the frontend design as well, then lead its development and deployment.
Decisions & trade-offs.
Asynchronous Job Processing: Anonymization jobs run through AWS SQS-backed workers rather than a synchronous request, so a large document doesn't block the interface. The frontend reflects job status instead of waiting on it.
Trust Over Speed: Built a review step instead of shipping the anonymized output directly, because a missed field is worse than a slower handoff: a side-by-side original/anonymized viewer, plus a replacements modal listing every substitution and flagging anything that should have been anonymized but wasn't.
Result.
Shipped and running in production, handling sensitive documents end to end from upload to anonymized download.
Stack.
TanStack Query for job status, SQS-backed workers for the async pipeline, Docker on AWS for deployment.
TypeScript
React
Shadcn UI
Tailwind CSS
React Hook Form
React Query
Zod
Fastify
Objection.js
Unit Testing
AWS
Terraform