Accountability infrastructure for AI decisions
Making AI decisions accountable before they cause harm.
I build accountability infrastructure for AI decisions — evidence traceability, real-time reconsideration controls, and human override pathways. Shipped production systems across EHR workflows, clinician communication, oncology, and care navigation.
Resume · Epic EHR · Doximity Dialer (110M+ clinician calls) · Director of Product at Transcarent · Andwise co-founder · Accountability infrastructure builder
Current focus AI safety and applied AI — Building verification infrastructure — systems that detect when clinical guidelines, policies, or underlying facts become outdated.
Case studies
Identity verification that made telehealth answering possible
Patients ignored calls from unknown numbers, blocking telehealth adoption. Dialer showed the doctor's office number — verified and HIPAA-safe.
Routing rules that made care navigation accountable
Led product work across value-based specialty-care and care-navigation programs (Surgery, Urgent Care, Behavioral Health, Oncology Care).
Review routing and escalation for physician-facing financial guidance
Co-founded Andwise; public pages corroborate founder and advisory-board context, while growth and funding metrics are founder-reported.
A safety framework for AI decisions
Three core tests before relying on an algorithmic decision in production: Is it explainable? Can someone challenge it? Can it be corrected when facts change? Read the framework →
Organizations I've worked with
Current work
Methods and research
Governance operating systems for AI safety — making decisions explainable, challengeable, and correctable in production.
Governance operating system
Ethotechnics
A governance methodology and operating system for making AI decisions explainable, challengeable, and correctable in production — with explicit ownership, escalation, and correction paths.
Explore method →
Specification
Mechanized Auditability Protocols for Algorithmic Systems
A technical specification for making non-deterministic model decisions verifiable and challengeable — with immutable ledgers, authorization states, and reconsideration loops.
Read specification →
AI and evaluation systems
If you're building AI systems that need accountability, evaluation, or correction infrastructure — I'd like to hear about it.