Built RESTful and WebSocket APIs enabling real-time fraud alerts, dispute notifications, and transaction review workflows for risk analysts.

Integrated Stripe Radar, Experian, and LexisNexis APIs to enhance fraud models with identity verification, device fingerprinting, and behavioral analytics.

Integrated ML-based fraud-scoring models into backend pipelines, enabling real-time risk prediction and improving high-risk detection accuracy by 14%.

Developed asynchronous microservices using Celery, RabbitMQ, and Kafka Streams, reducing fraud-analysis latency by 35%.

Built data pipelines for feature extraction and ML inference using Python, Kafka Streams, and PostgreSQL to support continuous model improvements.

Implemented automated dispute and chargeback pipelines using PostgreSQL, Redis, and batch processing, improving dispute win rates by 18%.

Designed and enforced OAuth2, RBAC, and service-level access controls to secure customer and fraud-risk data across distributed systems.

Built internal dashboards with React.js, TypeScript, and Material UI to visualize fraud scores and trends, reducing investigation time by 30%.

Containerized services using Docker and deployed on AWS ECS/EKS, implementing CloudWatch monitoring, S3 log archival, and anomaly alerts.

Automated infrastructure with Terraform and optimized GitLab CI/CD pipelines for reliable multi-environment deployments.

Developed unit, integration, and performance tests using Pytest, unittest, and Postman, achieving 95%+ test coverage.

Collaborated with Risk Science, Compliance, SRE, and Data Engineering to optimize fraud pipelines, improving response times by 22%.

Implemented observability with Prometheus, Grafana, and ELK Stack to detect fraud spikes, API anomalies, and ML model drift.

Architected and implemented event-driven systems using Apache Kafka, AWS Kinesis, and AWS SQS, processing millions of real-time events per day with low latency and high availability.

Built distributed data processing pipelines using PySpark and Spark SQL on Amazon EMR to generate terabyte-scale feature datasets, metrics, and analytical outputs.

Orchestrated batch and near-real-time workflows using Apache Airflow, implementing automated backfills, dependency management, failure recovery, and 99% SLA compliance.

Designed and optimized relational and NoSQL data models using PostgreSQL and MongoDB to support transactional and semi-structured workloads, improving query performance by 30–35%.

Developed RESTful APIs with proper versioning, authentication, and authorization, following secure API design and AWS IAM best practices.

Built internal web applications and dashboards using React and TypeScript to enable monitoring of system health, data pipelines, and business KPIs.

Integrated machine learning inference services into backend APIs to support personalization, optimization, and anomaly detection use cases.

Collaborated with data science and ML engineering teams to productionize machine learning models, focusing on feature pipelines, scalable inference, and performance monitoring.

Implemented AWS data lake architecture using Amazon S3, enabling analytics and reporting through Amazon Athena and Amazon Redshift over multi-terabyte historical datasets.

Containerized services using Docker and deployed them on Amazon EKS (Kubernetes), implementing auto-scaling, rolling deployments, and resilient system design, reducing deployment-related incidents by 25%.