Shipped 100+ AI agents and governed 30+ models under ERISA, SOX, HIPAA, and fair lending.
I translate AI's potential into operational reality — building the governance, execution, and measurement infrastructure that turns ambitious AI strategy into programs that are trustworthy, compliant, and built to deliver in regulated environments.
From brake to accelerator: I build responsible AI frameworks that let leadership move faster because the risk envelope is clear, defensible, and already regulator-tested.
Tie every AI dollar to the P&L — with investment theses, unit economics, and post-deployment measurement that finance teams can stand behind.
I've scaled AI teams from 15 to 300+. The bottleneck is never compute or models — it's literacy, incentives, and operating structure. I build the scaffolding that makes adoption inevitable.
I've seen where AI programs typically stall between pilot and production in financial services — and I know how to close that gap under ERISA, SOX, HIPAA, and fair lending.
Two decades building and governing AI in regulated financial services — from enterprise advisory to leading AI at a major credit union.
Lead enterprise AI strategy, agentic AI deployment, and model risk governance for one of the Northeast's leading credit unions.
AI transformation consulting for growth-stage companies and authored two books on causal inference and AI in financial services.
Drove enterprise AI/ML transformation — production GenAI, model risk, and MLOps at scale.
Built and led enterprise data engineering and analytics powering the firm's AI and reporting platforms.
Led enterprise advisory and AI/ML delivery for clients across financial services.
Peer-reviewed and preprint work at the intersection of causal ML, fairness, and large-scale data systems in financial services.
Develops a doubly-robust estimator applied to ~90,000 mortgage applications, revealing that 77% of the racial denial gap operates through financial mediators shaped by structural inequality.
Read on arXivHybrid model enhancing Legal-BERT through semantic similarity filtering. Achieves 93.4% F1 on 15,000 annotated legal documents.
Read on arXivNovel MDM algorithm achieving 90% accuracy on 10M+ records with 30% latency improvement, using PySpark and Databricks with Delta Lake.
Read on arXivJoin leaders getting weekly practical briefs on enterprise AI strategy, ROI frameworks, and what's actually working in regulated industries.
Why AI creates more hidden knowledge inside your organization, not less.
Read →Grounded, auditable document extraction built for regulated teams — credit unions and hospitals included.
Read →“It feels like it's working” isn't an ROI case — how to prove a pilot actually paid.
Read →Why a structured council of AI roles beats asking a single model what it thinks.
Read →On deploying autonomous agents securely as the fastest path to outsized ROI.
Read →How open-source causal inference is transforming financial risk assessment.
Read →For speaking, research collaborations, or trading notes on enterprise AI in regulated industries, send a few lines on what you're working on.