Projects
From regulatory AI at Goldman Sachs to personal ML experiments — a mix of what I build at work and what I build for myself.
GenAI Regulatory Analytics Platform
Spearheading an advanced generative AI platform that automates dynamic Federal Reserve regulatory report generation. Uses multi-agent LLM orchestration with specialized agents for data extraction, quantitative analysis, narrative generation, and compliance checking — dramatically reducing manual reporting overhead.
Liquidity Risk Compliance Platform
Architecting mission-critical infrastructure for computing FR2052a and Federal Reserve Board mandated liquidity risk metrics. Includes a full AWS cloud migration (ECS, Fargate, MWAA) and multi-stage Apache Airflow DAGs orchestrating end-to-end risk calculation pipelines with comprehensive error handling.
Internet Banking Microservices Platform
Designed and deployed 25+ production microservices for an internet banking platform serving 400+ credit unions at 5M+ transactions per day. Includes an LLM-powered Virtual Banking Assistant fine-tuned on customer financial data for spending analysis and personalized insights.
Real-Time Transaction Analytics Platform
Designed a platform to analyze high-velocity financial transaction data using Apache Kafka and Python, enhancing fraud detection capabilities by 40%. Built interactive React dashboards enabling financial institutions to monitor real-time transaction trends and anomaly signals.
Intelligent Credit Risk Assessment
Implemented a machine learning model for dynamic credit risk evaluation using TensorFlow and scikit-learn, improving classification accuracy by 30%. Integrated secure financial data APIs for real-time credit assessment with full compliance validation.
Predictive Account Monitoring Tool
Developed a forecasting tool for detecting unusual account activity using Python anomaly detection algorithms. Automated reporting pipelines with Google Cloud DataFlow for real-time alerting on flagged transactions, enabling proactive fraud response.