Product Manager - AML

Oscilar is building the most advanced AI Risk Decisioning™ Platform for banks, fintechs, and digitally native organizations to manage fraud, credit, and compliance risk. The Product Manager for AML will have full ownership of the AML and financial crimes suite, focusing on transaction monitoring, KYC/KYB, screening, case management, and SAR filing, while leveraging AI tools to enhance workflows and decision-making processes.

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
• Case Management. The workflow layer: alert routing, case assignment, escalation logic, analyst productivity tooling, QA/QC review, disposition tracking. You think about this as a human-in-the-loop system design problem - where humans add the most value and where AI can carry the load
• 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

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