╔══════════════════════════════════════════════════════════╗
║ building machines that understand how humans decide. ║
╚══════════════════════════════════════════════════════════╝
Independent Researcher · NLP & Behavioral Finance AI
Studying the gap between rational models and real investor behavior in the Indonesian retail market.
I build domain-specific NLP systems to detect cognitive biases in financial text — FOMO, loss aversion, confirmation bias.
Currently focused on:
- Cognitive bias detection in Indonesian retail investor discourse
- Behavioral finance meets computational linguistics
- Building lightweight, interpretable models that outperform large general-purpose LLMs on domain-specific tasks
| project | what it does | stack |
|---|---|---|
| cognifi | Rule-based NLP for real-time cognitive bias detection in Indonesian investor text. 94.7% accuracy, outperforms Gemini 3 Flash by 18.3pp. | Python, NLP |
| Sebastianisme | A thinking framework for challenging objective reality, rejecting dogma and systemic bias. | — |
