The Human Layer
Coming SoonWhat I've learned about human feedback from evaluating AI at scale. Why the bottleneck isn't the ML — it's us.
Writing
Occasional essays on AI alignment, content quality, and the human side of machine learning systems. More coming soon.
From Human Rights to AI Safety
Coming SoonEvidence verification, ground truth, and why my non-traditional path might actually matter.
What 'Quality' Means When You're Training a Model
Coming SoonThe problem of defining quality at scale. Consistency vs. context. How evaluators shape behavior.
Open Questions I'm Thinking About
These are problems I find genuinely interesting — not because I have answers, but because I don't:
- •How do we build evaluation rubrics that are consistent and context-sensitive?
- •What's lost when we scale human feedback through labeling platforms?
- •How should conversational AI behave differently for children vs. adults?
- •What can human rights verification methodology teach AI data integrity?
- •How do we handle cultural variation in quality judgments?
If you're working on any of these, I'd love to compare notes.