Every tool here has been used in a shipped system. Below is the engineering stack, compliance domain knowledge, and the specific context in which each was applied.
Orchestrating 5-agent AML/KYC investigation pipelines in SENTINEL — triage → evidence → reasoning → SAR drafting.
Training and serving deepfake detection, liveness check, and smart ID capture models for regulated KYC workflows.
Production REST APIs for KYC/KYB risk engine, credit scoring, and compliance microservices — serving real-time decisioning.
Graph database powering UBO detection and relationship mapping across 27,000+ PEP profiles in AfricaPEP.
Risk models for credit scoring with XDS bureau data — features, training, and automated decisioning output.
Citation-backed answers on KYC, AML, and STR regulations for LexAura's African regulatory Q&A interface.
Production model work and typed product interfaces.
Training, evaluation, inference, and model packaging.
Identity capture, liveness checks, OCR, and fraud detection.
Agentic workflows, retrieval, entity resolution, and compliance Q&A.
Product APIs, admin interfaces, dashboards, and workflow tools.
Operational storage, graph relationships, queues, and deployment.
Identity verification, document authentication, biometric matching, and onboarding automation.
Business verification, UBO detection, corporate registry checks, and entity resolution.
Employee screening, PEP checks, sanctions checks, adverse media, and ongoing monitoring.
Transaction monitoring, typology mapping, risk assessment, alert triage, and reporting.
Real-time thresholds, behavioral analytics, velocity checks, and geographic risk signals.
FATF recommendations, customer due diligence, enhanced due diligence, and audit trails.
AML/CFT recommendations and risk-based compliance principles.
OFAC, UN, watchlist screening, and ongoing monitoring patterns.
Deepfake, spoofing, replay, anomaly, and suspicious activity detection.
Evidence gathering, case summaries, risk explanations, and SAR drafting.