Nahrin Oda

Independent researcher & engineer. I study systems people depend on — starting with AI hiring — and build open tools grounded in the research: bias audit harnesses (Winnow), fine-tuned models with calibration, abstention, and fairness non-regression gates (inkling-hire, first in the inkling-* series built on Thinking Machines' Inkling). Interests: fairness & bias evaluation, LLM calibration, fine-tuning, synthetic data, forecasting. Open source at github.com/navox-labs — research first, build second, ship when it's real.