Lera Goulah

I've worked with SoftServe to evaluate AI behavior against intended product scenarios, clarify evaluation criteria with ML engineers, and surface reproducible failures and edge cases.

I provide evidence-based feedback that focuses on root causes, challenge assumptions about feature readiness, and help teams understand gaps between expected behavior and actual results.

As a freelance linguistic data specialist for DeepL, I contribute training data for LLM training and fine-tuning through linguistic evaluation, controlled text transformations, and style-transfer work.

My background as an editor and translator at Intext Translation Company strengthened my approach to semantic accuracy, pragmatic intent, consistency, quality standards, and ambiguous-case assessment.