Conversation Labeling – Ozmo Agentic AI Call Center Pipeline
TextClassification
In the Ozmo Agentic AI Call Center Pipeline project, I labeled call center conversation transcripts for downstream NLP analysis. I employed BERTopic for topic modeling and visualized results using UMAP. This labeling work enhanced sentiment analysis and model evaluation processes.
• Labeled call transcripts for topic and sentiment analysis
• Applied BERTopic and UMAP to categorize conversations
• Supported fine-tuning of local LLM for conversational AI
• Assessed model performance using labeled data.
In the Ozmo Agentic AI Call Center Pipeline project, I labeled call center conversation transcripts for downstream NLP analysis. I employed BERTopic for topic modeling and visualized results using UMAP. This labeling work enhanced sentiment analysis and model evaluation processes. • Labeled call transcripts for topic and sentiment analysis • Applied BERTopic and UMAP to categorize conversations • Supported fine-tuning of local LLM for conversational AI • Assessed model performance using labeled data.
2026 - Present