At Purlin Co., I validate production ML models for document classification, entity extraction, signature detection, image classification, and similarity search. I run regression tests on unseen real-world data and track precision, recall, and F1 to find performance drops.
I investigate incorrect predictions with ML engineers, tracing issues to model behavior, class imbalance, or labeling errors. I also create synthetic documents for edge cases and use Langfuse and Promptfoo for prompt versioning and evaluation.
Across Purlin Co., Siemens Digital Industries Software, Digitain, and AOD Space, I’ve tested web, mobile, API, and ML-powered applications. I’ve automated tests with Python, Pytest, and Cypress, and assessed load and stress behavior with Locust and JMeter.
