Prashant Sharma

Language Data Specialist with 1.5+ years of hands-on experience in LLM evaluation, human-in-the-loop workflows, AI data validation, and multilingual language quality assurance. Experienced in designing and improving annotation workflows for NLP and speech recognition systems. Strong background in data preprocessing, statistical analysis, and quality metrics.

Associate Prompt Engineer at Randstad Pvt Ltd (2024-08 – 2025-05)

• Designed and executed human-in-the-loop evaluation workflows for Large Language Models (LLMs), improving response accuracy and linguistic precision by ~20%.
• Evaluated AI-generated outputs for contextual correctness, ambiguity detection, hallucination identification, and intent alignment.
• Flagged systematic model errors and contributed insights to improve upstream data collection and preprocessing pipelines.
• Collaborated cross-functionally with AI engineers and QA teams to refine evaluation metrics and improve workflow efficiency.

Data and Annotation Engineer at P&R Groups (2025-07 – 2026-04)

• Designed and maintained data pipelines to support NLP and ML workflows, ensuring reliable data ingestion, transformation, and storage.
• Implemented data validation and quality assurance checks to improve dataset integrity for downstream AI/ML models.
• Optimized data transformation processes to enhance efficiency, consistency, and reproducibility.
• Documented end-to-end data processes to ensure scalability and cross-team usability.

Diploma in Data & Generative AI – Ducat Institute (2023 – 2024)

Bachelors in Computer Applications (BCA) – Chaudhary Charan Singh University (2019 – 2022)