Sarvesh

AI Analyst

Innopulse Consulting GmbH • Fulltime

Dec 2025 - Present • 7 mos

As an AI Analyst at Innodata, I work on improving the performance, quality, and reliability of Large Language Models (LLMs) through prompt engineering, data annotation, model evaluation, and quality assurance processes. My responsibilities include designing and optimizing prompts, evaluating AI-generated responses against predefined quality standards, and supporting Reinforcement Learning from Human Feedback (RLHF) workflows. I collaborate with AI researchers, data scientists, and QA teams to identify model weaknesses, improve response accuracy, and ensure consistent output quality across diverse use cases. In addition to AI-focused work, I analyze operational datasets using SQL and Excel to uncover performance trends, validate data integrity, and support model fine-tuning decisions. I also contribute to workflow automation by developing data validation and reporting processes that improve efficiency and reduce errors. Through a combination of analytical thinking, attention to detail, and AI expertise, I help deliver high-quality training data and evaluation frameworks that enhance overall model performance and business outcomes.