Synthetic Sentience

Jul 2025 - Aug 2025

Description: A proof-of-concept chatbot that uses Hugging Face Transformers for emotion classification of user input, dynamically adjusting its own “mood” in response. Detected emotions were amplified (e.g., +0.25 intensity) and decayed over time per message, allowing the bot’s tone and phrasing to shift in a way that loosely mirrored the user’s emotional state. The goal was to explore how machine learning can drive emotionally adaptive dialogue.

Technologies:

Python

Next.js

Hugging Face

FastAPI

Gemini

VSCode