Tariq Nwachukwu

Hierarchical spans and explicit context propagation enable accountability across agent boundaries.

WebShop tests only product search, missing five critical e-commerce behaviors.

Automated testing catches AI agents answering the wrong question while appearing to succeed.

Structuring how AI systems label multi-step task data at scale.

LLM judges hold up on text tasks but fail in agent pipelines.

Most LLM monitoring logs outputs without judging them, and that's the gap.