Sambit Barik

I’m a results-driven Senior SDET / GenAI Test Engineer with 6+ years of experience delivering quality at scale for AI-driven products. I focus on LLM evaluation, RAG pipeline testing, and end-to-end automation for conversational AI, Text-to-SQL, and OCR systems.

I build robust test frameworks that reduce manual effort by up to 80%, including Playwright, Pytest, and Selenium. At Fractal Analytics, I architected an LLM-as-Judge automation framework using Postman + Claude API to evaluate response quality, cutting manual review cycles by ~60%.

I also ship repeatable validation into CI/CD pipelines so teams can make faster, safer releases. I integrated LLM evaluation metrics (Langfuse, Ragas, DeepEval) for real-time visibility into model drift and regression, and automated ETL workflow validation on Databricks with Pytest to protect zero data-loss SLAs.

My approach bridges model validation with business outcomes using domain-specific vocabularies, few-shot learning, and semantic scoring. I’m passionate about responsible AI—ensuring accuracy, fairness, and production-readiness for enterprise-scale deployments.