Raghavendar Ramachandran

Focused on training, evaluating, and stress-testing multimodal AI systems (text, audio, image, and video) to identify edge cases, enhance reasoning, and improve content generation accuracy. Specialized in Large Language Model (LLM) training and evaluation, improving factual accuracy, coherence, and contextual reliability across diverse content domains. Designed and executed stress and edge-case testing frameworks to intentionally break models, uncover limitations, and deliver insights for fine-tuning and retraining. Served as Human-in-the-Loop (HITL) to refine AI outputs through prompt optimization, fact verification, linguistic enhancement, and tone alignment. Evaluated video understanding models for accuracy in information extraction, temporal reasoning, and contextual comprehension. Assessed audio and image-based models, generating prompts and grading outputs for realism, fluency, and semantic relevance, while collaborating with research teams for improvement. Conducted detailed data validation and content benchmarking to ensure linguistic precision, factual consistency, and adherence to quality standards. Maintained 100% quality scores in all evaluation cycles. Awarded the Extra Mile Award for exceptional ownership and proactive problem-solving. Recognized as Top Performer of the Month (October 2024) for excellence in AI model testing and reliability improvement.

Created and managed social media strategies and content for Dream Tamilnadu and Idea Pattarai projects. Increased social media reach by 30%, resulting in an average gain of 4,000+ followers per platform. Organized and coordinated workshops, ensuring seamless stakeholder communication and event execution. Wrote, proofread, and edited scripts for educational videos and entrepreneurship curriculum projects.

Assisted in daily show production, contributing to content ideation, script editing, and outdoor shoot coordination. Supported producers with proofreading and content development aligned with brand guidelines.