Nabla-X Insights
Intelligence for the regulated world
Research, guides, and perspectives on AI in healthcare, finance, and science.
All articles
The Hidden Cost of Bad Data in Financial Services
A single corrupted field in a risk model can cascade into millions in mis-priced exposure. We analyzed 50 production data incidents to understand where financial data quality breaks down — and how to catch it earlier.
Beyond SHAP: A Practical Guide to Root Cause Analysis for ML Models
SHAP values tell you which features mattered. But they don't tell you why a specific prediction was wrong, or how to fix it. This guide covers the techniques we use in Decision Intelligence to go deeper.
Building HIPAA-Compliant AI Infrastructure: What We Learned
After two years of deploying AI in hospital systems, we've accumulated a playbook for infrastructure that actually passes audit. This is the guide we wish we'd had on day one.
AI Observability: What It Actually Means for Production Systems
Observability for software means logs, metrics, and traces. For AI systems, it's more complex. We define what meaningful AI observability looks like when models are making consequential decisions.