Personality Based Chatbot with Retrieval Augmentation
Built a real-time vision pipeline for multi-camera fusion, detection, and tracking in complex traffic scenarios.
Highlights
- Built a persona-driven chatbot that imitates distinct TV-character voices (e.g., Chandler and Penny) using an end-to-end neural dialogue generator (no hand-written rules).
- Curated and transformed episode screenplay/dialogue datasets into multi-turn training samples via regex/string-cleaning and context-window construction (n prior turns → target character reply).
- Fine-tuned DialoGPT for open-domain conversations; the bot produced character-consistent, fluent responses and was evaluated with perplexity (≈0.75–0.82 range depending on run/setup).
Demo
Next: embed a Hugging Face Space here (or custom inference UI).