JAS/Work/BigBossCorp
←Back to Work

Oscilar

Designed and developed backend services using Java (Spring Boot) and Python for an AI Risk Decisioning Platform, enabling real-time fraud, credit, and compliance decisioning with Kafka event streaming and ML model integration.

Tech Stack

java iconJava
spring iconSpring Boot
python iconPython
fast-api iconFastAPI
kafka iconApache Kafka
reactjs iconReact
typescript iconTypeScript
nextjs iconNextJS
postgresql iconPostgreSQL
redis iconRedis
pytorch iconPyTorch
xgboost iconXGBoost
scikit-learn iconScikit-learn
kubernetes iconKubernetes
aws iconAWS
ci-cd iconCI/CD
jenkins iconJenkins

Roles

Senior Full Stack/AI Engineer

02/2023 – PresentRemote
  • Designed and developed backend services using Java (Spring Boot) and Python for Oscilar AI Risk Decisioning Platform, enabling real-time fraud, credit, and compliance decisioning across financial transactions.
  • Built low-latency decisioning services, reducing average response time to under 100ms for transaction risk evaluation across high-volume streaming pipelines.
  • Implemented event-driven architecture using Kafka, processing thousands of events per second for real-time ingestion and decisioning workflows.
  • Designed and implemented RESTful APIs for transaction processing and risk evaluation, enabling reliable communication between distributed microservices.
  • Designed and optimized PostgreSQL schemas for transaction and user data, improving query performance and ensuring consistency for high-volume financial workloads.
  • Implemented caching strategies using Redis, reducing data retrieval latency by ~40% and improving performance for real-time decisioning services.
  • Developed Python-based services for machine learning and AI workflows using FastAPI, enabling scalable and low-latency model integration.
  • Integrated machine learning models using PyTorch, XGBoost, and Scikit-learn, improving fraud detection accuracy and reducing false positives by ~20%.
  • Designed microservices-based architecture deployed on Kubernetes, improving scalability, resilience, and supporting high availability across distributed services.
  • Built internal dashboards using React and TypeScript, improving analyst efficiency and reducing manual review time for flagged transactions.
  • Developed LLM-powered features for case summarization and risk analysis, improving investigation efficiency and reducing manual effort.
  • Built retrieval-based pipelines combining real-time and historical data, improving contextual accuracy and explainability of AI-generated insights.
  • Leveraged AWS infrastructure for scalable deployment, monitoring, and high availability of real-time decisioning systems.
  • Implemented automated testing and CI/CD pipelines, reducing deployment time by ~35% and improving release stability across backend services.
  • Leveraged AI-assisted development tools to improve code quality and reduce debugging time across backend and frontend workflows.

Key Projects

  • Oscilar's AI Risk Decisioning Platform: Real-time fraud and risk decisioning system that helps financial institutions detect anomalies, reduce false positives, and make faster, smarter approvals using ML-powered insights.