Personal project · NLP experiment
Making sentiment, emotion and sarcasm visible through a playful Discord bot.
A Discord bot that reads a message, detects sentiment, emotion and sarcasm, and replies with a suitable emoji.

01 · Context
The problem
I wanted to explore how a model trained on 1.2 billion tweets could interpret the tone of everyday team messages and turn its prediction into a lightweight social interaction.
02 · Responsibility
What I was responsible for
- 01Connected the trained language model to Discord messages.
- 02Mapped sentiment, emotion and sarcasm predictions to suitable emoji reactions.
- 03Packaged the experiment as a bot colleagues could interact with.
- 04Documented both the project and its source code publicly.
03 · Evidence
What happened
The bot was deployed on an office server for colleagues to try, and the implementation remains linked on GitHub.