Product Manager - AML

Oscilar is building the most advanced AI Risk Decisioning™ Platform to help banks and fintechs manage fraud and compliance risk. The Product Manager for AML will own the product suite related to anti-money laundering and financial crimes, focusing on transaction monitoring, KYC/KYB, screening, and regulatory reporting.

Responsibilities:

• Transaction Monitoring. The rules engine and ML model layer that generates alerts - scenario design, threshold calibration, typology coverage, backtesting against historical transaction data. You'll define what the system flags, why, and how those decisions evolve as customer behavior and fraud typologies shift. You understand the difference between a rules-based velocity check and a graph-based network detection model, and you have opinions on when each is right
• KYC / KYB. Onboarding risk workflows: identity verification, document checks, beneficial ownership, ongoing monitoring triggers, and risk re-scoring. You know the data sources that power these decisions - bureau data, device signals, open banking, registry data - and how to orchestrate them dynamically based on risk tier rather than applying a one-size waterfall
• Screening. Sanctions (OFAC, UN, EU), PEPs, adverse media, and negative news. You understand the matching problem - why fuzzy matching on names produces the false positive rates it does, and what it takes to make screening defensible to an examiner without generating alert volumes that bury the ops team
• SAR Filing & Regulatory Reporting. FinCEN SAR workflow end-to-end: narrative generation, supporting documentation, filing status, audit trail. You've seen what regulators look for in an MRA and you build product that holds up under that scrutiny
• AML Data Strategy. The data model that powers all of the above: transaction graphs, entity resolution, behavioral baselines, consortium signals, model feature pipelines. You can speak fluently to a data engineer about what data the models need and why, and to a compliance officer about what the model's output means and how it was derived
• You are an AI-native PM. You use AI tools throughout your entire workflow — not as a novelty, but as a force multiplier that lets you do in a day what used to take a week
• For discovery: You use AI to synthesize call transcripts, cluster themes across interviews, scan regulatory guidance and examiner findings, and surface competitive product gaps faster than any manual process. You come into customer conversations already knowing what the literature says - so you can spend the time extracting what the literature doesn't
• For building: You use AI to generate first-draft PRDs, stress-test your own specs, write and iterate on user stories, prototype UI flows, and pressure-test edge cases before engineering touches them. Your specs are tighter and your stakeholder review cycles are shorter as a result
• For design: You use AI to explore multiple UX directions quickly and give the design team a sharper brief. You're not a designer - but you use AI tools well enough that you show up with something specific rather than a blank page
• For analysis: You use AI to query product and operational data, interpret model outputs, and build the evidence base for roadmap decisions. You don't wait for a data analyst to unblock you

Requirements:

• Full-stack AML product depth. You've shipped product across at least two of the five areas above - in production, with real users, under real regulatory scrutiny. You've sat in a room where an examiner asked a question about your system and understood exactly what was at stake. You know what a BSA officer worries about before an exam, and you build product that addresses it
• AML strategy knowledge. You understand how AML programs are designed and calibrated, not just executed. You know why scenario thresholds drift over time and how to fix it. You know the difference between a structuring typology and a layering typology and have an opinion on which detection approach fits each. You understand model drift, backtesting methodology, and what a model validation report covers. You can engage a Chief Compliance Officer on their AML program design and add something to the conversation - not just listen
• Technical and data fluency. Your engineers trust your specs. You can read a data schema, write a query to validate a hypothesis, and spot when an ML model is being misapplied. You understand the explainability problem in AI-assisted compliance decisions - not as an abstract concern, but as a concrete product constraint - and you have a position on how to solve it
• Regulatory literacy. BSA, FinCEN, OFAC, FATF. You understand what examiners are looking for, what triggers an MRA, and how to build a defensible audit trail. You've worked in an environment where regulatory guidance was unclear, evolving, or in conflict with shipping speed - and you made the call anyway
• Bias toward speed. You use AI tools to move faster, test sooner, and learn earlier. You do in a week what takes a month elsewhere