Tag: Engineering Rigor

One pass rate can hide a zero: introducing the Agent Reliability Kit

An adversarial suite can pass at 90% while one kind of attack gets through every time. We've open-sourced the eval harness and scorecard we use to catch that.

Pushpendra Singh

Pushpendra Singh

Oct 28, 2026

11 mins read
AI Evals Agentic AI AI Safety Open Source

Evals before Features: Shipping Reliable Agentic AI Systems

Reliable Agentic AI starts with a clear way to measure quality. This guide shows how to define representative evals, track what matters, and use the results to guide development from the first feature onward.

Pushpendra Singh

Pushpendra Singh

Jul 22, 2026

16 mins read
AI Evals LLM Agentic AI Quality Engineering