TLDR The redaction hypothesis tested in a controlled environment showed Presidio redacted 97% of the target name in the dataset. The [Redact, NER, Analyze] pattern was implemented against “Moby Dick” to remove all instances of Ahab. The outcome suggests a valid testing pipeline, easily abstracted to sparknlp for exploration at scale…

TLDR Serverless containers provisioned by Fargate and triggered by Lambda good. PyODBC deployed on Lambda bad. Use Fargate for long-running jobs or jobs that are well supported as containers and minimize CICD build complexity. PyODBC vs Lambda In the very recent past I was wrapping up what I thought to be the more…

In this demo, I will be deploying Microsoft Presidio to AWS. This is not a production-grade deployment. This is not a production-grade deployment. This is not a production-grade deployment. Why? Why is this not production-grade? It’s not autoscaling! I’m setting up docker containers without restart policies on an EC2; terrible…

TLDR Skip to the bottom. The code is down there. What is Presidio? Presidio is an anonymization tool created by Microsoft and open-sourced by some sweet unicorn in those hallowed halls. A few minutes spent in your favorite search engine poking around for redaction services or redaction libraries will yield a great big ole…