Darius Griffin

Nazca

AI enablement and compliance advisory for regulated institutions.

I build and deploy agent-driven compliance systems for teams operating inside regulated financial environments: adverse media, sanctions review, case triage, policy inventory, document translation, and compliance operations that usually get buried under volume.

The model is simple: agents do the repetitive work; humans stay accountable for the consequential calls.

This is not theory. The work is live, launched, and earning. It is helping regulated financial platforms build compliance infrastructure, ship new products, and bring AI into real control environments.

The first call when AI meets a regulated balance sheet.

Dates being scheduled

Talks & workshops on AI in compliance operations, from someone who deploys it.

Nazca

Nazca is a compliance OS powered by agents, the first proof of the AI enablement and compliance advisory thesis.

An adverse-media screening agent. A case-management system. Policy-and-procedure inventory maintenance. Document translation. Compliance operations rebuilt around agentic volume work and human accountability.

Antler Spring ’25

Antler's Spring '25 cohort is where I went all the way down the AI rabbit hole: rapid prototyping in Replit, automations in n8n, agent development, validation, testing, and knowledge graphs.

I came out with the realization that reorders everything after it: the barrier is gone.

Technology used to be routed through a small class of people with a specific technical skill set. Not necessarily the most creative people. Not necessarily the smartest. Not necessarily the ones closest to the problem.

Now the work goes to whoever understands the problem best.

Beam acq. by Modern Treasury

Built the program from zero: money-transmitter licensing across states, transaction monitoring, sanctions screening, policy, and controls, laying the foundations for a clean acquisition.

That chapter opened a new lane for me in stablecoin payment rails and orchestration. It also taught me customer-concentration risk the real way. I prepped the government-contracting path through Dcode's accelerator, then learned the harder operating lesson: when the fit is wrong and the upside is not truly worth the time, leave cleanly and move on.

I did.

OKX

Global KYC across EMEA, APAC, and the Americas: complex entity structures, distributed teams, staff upskilling, and product management across time zones. It was also my first heavy use of AI inside the work itself.

The lesson that stuck: a program built correctly at the start becomes an asset. Built poorly, compliance becomes a permanent cost center carrying regulatory risk forever.

Great seat. Great pay. But I kept missing the asymmetric bet: that early-stage-to-$3.5B feeling.

BlockFi

One of the first compliance hires. Rode it to Director. Fifty-plus people. Expansion across five countries. Seats at Chainalysis's Client Advisory Board and TRUST's committees. Then call it what it was: the company took on too much risk, the winter came, and it ended in bankruptcy.

These same years taught me both sides of the startup equation: join early, because that is where the win lives; hold cash, because paper is not money.

But the lesson that stung most was watching decisions that needed making sit outside my influence.

Master of your own fate is not a slogan. It is the reason to found.

The sabbatical

A few years traveling with my young kids: the most formative money I have ever spent, for them and for me.

Somewhere between countries, I got technical for the first time: forked chains, modded games, wrote whitepapers. I came home wanting to build, not just review what other people built.

Coinbase

I had been building mining rigs at home to understand what crypto actually was. Then I found out Coinbase was hiring and applied, not fully knowing what I was walking into.

First startup. Brilliant people. Push the envelope. Move fast. Figure it out. Execute.

A whole new world. The bug bit and never let go.

Citigroup → OneMain

A year in Citi's management training program, then its lending subsidiary: branch operations, running and scaling my own branches, growing production from $14M to $20M.

Full Glengarry sales culture. I kept a low-delinquency book anyway.

Watching precomputed-interest loans become some people's only option sent me down a rabbit hole: CDFIs, community lending, and an old idea that never let go of me. Susus, the lending circles communities run on trust.

RadioShack & JAM Catering

Worked my way through Temple catering weddings for JAM and managing a RadioShack.

Raspberry Pis. Arduinos. The smartphone wave. The first taste of entrepreneurship: running the store like it was mine and helping people write business plans at the SBDC on Cecil B. Moore.

I learned to sell. I learned to fix. And I figured out early the desire to create.