Modular Agentic Dialogue Engine (Conversation Model) – AI Training/Data Labeling
TextFine Tuning
This role involved fine-tuning pre-trained language models for intent and command classification. Semantic memory retrieval using embeddings and persona-based response generation were key components. The experience also included designing logic for mood state handling based on LLM-extracted information.
• Fine-tuned DistilBERT models on intent and command classification datasets.
• Developed embedding-based memory systems for text retrieval.
• Implemented evaluation and fallback logic in dialogue engine workflows.
• Managed persona and emotion tracking for natural conversational AI.
This role involved fine-tuning pre-trained language models for intent and command classification. Semantic memory retrieval using embeddings and persona-based response generation were key components. The experience also included designing logic for mood state handling based on LLM-extracted information. • Fine-tuned DistilBERT models on intent and command classification datasets. • Developed embedding-based memory systems for text retrieval. • Implemented evaluation and fallback logic in dialogue engine workflows. • Managed persona and emotion tracking for natural conversational AI.
2026 - Present