Published Research
Original research and analysis on AI strategy, agentic systems, and enterprise AI adoption.
How Do You Know When an AI Is Right?
A practical guide for product, engineering, and business leaders choosing among outcome checks, process supervision, automated feedback, and verifier ensembles.
Viola Cao
How Do You Build an AI Verifier That Works?
An engineering blueprint for candidate reranking, outcome and process reward models, scalable supervision, weak-verifier ensembles, calibration, and monitoring.
Viola Cao
The blind spot in every AI eval framework
RAGAS, DeepEval, LangSmith, TruLens — mature frameworks, genuinely useful. But they were built for RAG. This is what they systematically miss, why it matters at 97M+ monthly MCP downloads, and what the research landscape confirms.
Viola Cao
APEX: Agentic Pipeline EXecution Diagnostic Framework
19 failure modes across 4 layers of the tool execution pipeline — and three evaluation primitives to close the gap. The map your eval framework was never built to read, with the solution attached.
Viola Cao