Prashansa Singh

I’m an AI Specialist and Annotator with 3+ years’ experience in advanced AI training workflows, including Supervised Fine-Tuning (SFT) and RLHF. I focus on prompt curation, annotation quality, and model evaluation to raise real-world model performance.

As an Expert Prompt Curator (Jan 2025–Present), I curate and assess challenging prompts and model queries for benchmarking. I review interaction logs, evaluate rewritten queries for fidelity to user intent, and flag sensitive content in line with safety guidelines, while providing evidence-based rationales aligned to rubrics.

Previously, as an Annotator (Feb 2023–Dec 2024), I produced high-accuracy annotation datasets (95%+ accuracy) for complex gaming AI evaluation. I led and mentored teams, improved productivity by 20% through workflow enhancements and training, and refined annotation pipelines to reduce revision rates through proactive feedback handling.

Earlier roles as an AI Trainer (Feb 2022–Feb 2023; Dec 2020–Feb 2022) strengthened my ability to translate model outputs into actionable improvements. I developed personalized AI models that boosted engagement by 25% and improved diagnostic precision by 40%, and I led quality control initiatives while authoring pipelines to support data generation for training and deployment.