At INNODATA INC, I perform multimodal annotation across text, images, speech, and video, and review AI-generated outputs using project rubrics. My work includes speech aesthetics evaluation, speaker-interval annotation, and visual grounding tasks.
At NVIDIA (via RANDSTAD TECHNOLOGIES PVT LTD), I annotated LiDAR point-cloud and multimodal data for autonomous vehicle and AI training workflows, and evaluated LLM responses for factual accuracy, relevance, instruction adherence, and safety. I also developed an LLM evaluation framework to assess response accuracy, relevance, clarity, and instruction adherence.
