Krishnendu Dubey

I’m a Data Annotation Operations Lead with 7+ years of experience leading large-scale AI data programs, annotation operations, model evaluation, and quality governance across NLP, Computer Vision, Conversational AI, and multi-modal datasets.

I’ve owned end-to-end annotation lifecycle management—task design, quality audits, reviewer calibration, reporting, and stakeholder communication—while building frameworks like Goal Success Rate (GSR) to improve evaluation consistency and business alignment. I also automate reporting and workflows with Python, cutting analysis turnaround time by 50% and reducing manual QA effort by 40% through process optimization and bug-tracking workflow improvements.

Across roles at Providence India, Apple India, and Amazon India, I’ve evaluated AI outputs (including RAG, GPT, and Computer Vision) using precision, recall, and F1-score metrics, built Tableau dashboards for throughput and operational KPIs, and improved audit turnaround time by 35% via reviewer onboarding and QA process redesign. I bring a pragmatic, quality-first mindset—standardizing guidelines, schemas, and operations to deliver reliable datasets for AI training, RLHF workflows, and model evaluation.