Sajag Swami
सजग स्वामीサジャグ・スワミ
Statistics, applied LLM research, and experiments
Tokyo, Japan.
I like taking things I learn and trying them out in the real world.
Sometimes that means building something. Sometimes it means collecting a little data. Sometimes it means reading a book and then noticing the same idea somewhere completely unexpected.
I'm particularly interested in statistics, causal inference, and understanding how AI systems behave in the real world, and finding useful applications for them.
Current interests
Causal Inference
I like turning everyday observations into questions that can actually be tested.
Applied LLM research
I'm interested in evaluation, small language models, and what model performance really tells us outside a benchmark.
Self-experiments
I have a habit of turning mildly unnecessary questions into datasets.
Professionally
I work at Mizuho, where I evaluate enterprise ML and LLM applications and think about what evidence is actually enough to trust a model.
What exactly are we measuring?
Elsewhere
- Scholar
- Google Scholar
- GitHub
- github.com/SunTzunami
- linkedin.com/in/s824s
- Medium
- medium.com/@sajagswami
- X
- @Swami824
- sajagtsunami@gmail.com