Eshmael Chava

I’m a Senior QA Manager and AI data quality leader with 6+ years of progressive experience delivering production-grade training datasets for Machine Learning, Computer Vision, NLP, and Speech systems. I focus on scaling QA operations for fully remote teams and translating data quality into measurable downstream model performance. In my current role, I lead QA operations for 60–120 globally distributed annotators and reviewers. I reduced annotation error rates by 32% and increased first-pass acceptance to 98%+. I also built audit frameworks covering 2M+ assets per quarter across CV, NLP, and speech datasets, improving inter-annotator agreement from 0.71 to 0.91 and delivering 97–100% SLA compliance across enterprise AI projects. Previously, as a data annotation expert and quality reviewer, I validated 5M+ labeled assets across multimodal datasets. I reduced dataset rework cycles by 35% through edge-case resolution and mentored annotators to improve individual accuracy by 18–25%. I also served as a QA team lead for remote AI data operations, managing 25–40 annotators and improving overall annotation accuracy from 85% to 95%. I bring hands-on expertise across annotation audits, guideline development, and data governance—covering SLA & KPI management, remote team leadership, and inter-annotator agreement tuning. My work spans computer vision (object detection, segmentation, keypoints), NLP & LLMs (NER, intent classification, sentiment analysis), and speech AI (transcription and speaker labeling), supported by tools like Labelbox, CVAT, SuperAnnotate, and Jira.