I build the platform, governance, and observability substrate that lets a regulated business ship AI automation on top of the systems it already operates — reviewed, drafts-only, and never touching the system of record. Anyone can demo an agent. The hard part is the substrate.
- Principal/staff AI engineer and engineering leader, ~20 years.
- Founded the AI delivery function at a regulated commercial real estate lender: a skill marketplace, a publish-and-audit governance gate, and observability with cost-per-skill. Full write-up: davidveksler.com/david/ai-strategy.html
- Drive agentic engineering at enterprise scale at Antech — Claude Code / Copilot workflows, ~4x developer velocity.
- Deep regulated-fintech and digital-asset background: CCSS Level 3 custody, $100M+ TVL, on-chain recovery tooling across 20+ EVM networks.
Stack: C#/.NET, Python, TypeScript/React, Azure, AWS, SQL, OpenTelemetry. Frontier-model + RAG over fine-tuning by default.
I don't train models and I'm not a prompt engineer. I get frontier models into production where the regulatory and security bar is high.
Reach me: [email protected] · LinkedIn · Substack





