Financial Crime Analyst

tl;dr: Sphinx is rebuilding how compliance gets done. We're hiring experienced AML analysts to work hand-in-hand with our AI agents: running real casework at banks and fintechs, shaping the system as you go. If AML has become second nature to you, and you want to define how the future of compliance looks, we should talk.

Sphinx builds AI analysts that automate AML, KYC, KYB, and transaction monitoring for banks and fintechs. Our agents work inside browsers and internal systems just like human analysts. They review alerts, investigate entities, file reports, and explain their reasoning.

Sphinx Frontline is how that mission shows up on the ground. Instead of just selling software, we run the operation, embedding our analysts and AI agents directly into customer workflows to handle compliance end-to-end. An AI-native compliance team-as-a-service.

That means our analysts aren't just working cases. They're answering the questions that decide how modern compliance actually functions: Half the alerts your team clears are noise, and you know it. So which signals actually matter? Where does an agent see things a human would miss, and where is it the other way around? And what does a financial criminal do that ten years of alert logic still misses? If those questions sound like the interesting part of the job, you're in the right place.

1. Rebuild compliance from the ground up: Most of the workflows that exist today were designed for humans clicking through screens. You'll get to ask what they should look like when an AI agent is doing the work, and then build that.

2. Your judgment becomes the system: Every escalation rule, every edge case, every "this shouldn't have been auto-cleared" instinct gets encoded into how the agents reason. You're not feeding a queue. You're writing the playbook the next generation of compliance runs on.

3. Hard problems, not busywork: When an agent flags something to you, it's because the routine logic ran out. You're working the cases that genuinely require judgment, and figuring out how to teach the system to handle the next one like it.

4. Real variety, real stakes: Frontline serves banks, fintechs, and payment companies across multiple jurisdictions. Every client surfaces different patterns, different regulators, different edge cases, and every one of them sharpens the system.

5. Define a new category: AI-native managed services is a new model for the industry. You'll be one of the first people to prove it works, and set the bar for everyone who follows.

• Real compliance casework. Investigate alerts, adjudicate transaction monitoring flags, conduct KYC/KYB reviews, prepare SARs and CTRs, and run screening investigations. The full spectrum of BSA/AML work across our client base.

• Working cases with the agents. You'll work alongside the agents on live casework, pressure-testing their reasoning, catching where they're wrong, and making sure every output meets the client's regulatory standards.

• Defining where humans belong in the loop. Which cases need a human eye, and why? What signals should trigger an escalation? You'll be the one designing those rules, turning instinct into criteria the agents can act on.

• Rewriting the playbook. When you spot a recurring false positive, a missed signal, or a jurisdictional quirk that the SOP doesn't account for, you'll fix it at the source. Updating the workflow, retraining the agent, and validating the change against historical cases.

• Real compliance experience: 2+ years in AML, BSA, KYC/KYB, transaction monitoring, or financial crime investigations at a bank, fintech, MSB, or consulting firm. You've worked real cases, not just studied the regulations.

• The workflows are second nature: Alert adjudication, SAR/CTR filing, screening investigations, EDD, document review. You've done it all and can do it without hand-holding.

• You think in systems: You've looked at a compliance workflow and thought "this is broken, and here's how I'd fix it." You want to redesign the work, not just execute it.

• You want to work with the agents, not around them: You don't need to be a developer, but you should be the kind of person who pokes at a new tool to understand how it works, and gets frustrated when a workflow is dumber than it needs to be.

• Strong judgment: You know when something looks wrong, when a case needs escalation, and when an AI output needs a second look. And you can articulate why, so the system can learn from it.

• You hold the line on accuracy: A missed SAR or a sloppy narrative isn't just a mistake. It's something a regulator will see. You know that, and you bring the kind of discipline that holds up under audit. The agents will move fast; your job is making sure fast doesn't mean wrong.

• You can make the complex legible: Whether it's a SAR narrative, a reasoning trace the agent needs to learn from, or an explanation for a regulator, you can take a messy investigation and turn it into something clear, defensible, and useful.

Frontline is remote-first. Work from anywhere.