A.I. THATHOLDS UP.
Selected work
Three studies / 01—03
Principles
How I think about systems.
- 01
Evaluation Harness
Can we prove it works?
- 02
Observability
Can we reconstruct what happened?
- 03
Faithfulness
Is what it says supported by evidence?
- 04
Idempotency
Can repetition happen safely?
- 05
Uncertainty Quantification
Do we understand the limits of what we know?
- 06
Causal Inference
Do we know what actually changes outcomes?
About
Applied AI,with an analyst’s skepticism.
I’m Erik Rice, an applied AI scientist and analytics engineer based in Chicago. My background is in data science, forecasting, and analytics in high-stakes public-health environments; my current focus is building AI systems that can be evaluated, traced, constrained, and trusted. I care less about whether a system can produce an impressive answer than whether its behavior can be explained, tested, and defended when the conditions change.