Data Engineer, but in my free time I'm obsessing passionately exploring AI interpretability.
- Mechanistic interpretability of neural networks
- Understanding how language models represent and process information
- Building tools for AI safety and alignment research
- TEMPO: Novel approach to LLM inference exploring parallel token processing for interpretability
- Vardon: Pathfinder 1e character engine (Work in Progress)
- AONPRD-Parse: High-performance async data pipeline demonstrating engineering expertise
I believe understanding how AI systems work internally is crucial for building safe, aligned AI. My work bridges practical engineering with theoretical understanding of machine learning principles.