• Built and scaled a fraud detection platform processing 4M+ daily transactions, using XGBoost and LightGBM models served via FastAPI,

helping reduce fraudulent approvals by 34% within the first quarter.

• Developed a real-time data pipeline with Apache Kafka and AWS Kinesis to capture and process transaction events with low latency,

enabling live fraud scoring through AWS SageMaker models.

• Created a React.js dashboard with Redux and D3.js visualizations, allowing analysts to review flagged transactions, adjust risk rules,

and investigate patterns more efficiently, reducing investigation time by 40%+.

• Implemented an LLM-based case summarization feature using LangChain and OpenAI APIs, generating clear fraud narratives from raw

data and reducing manual documentation effort for analysts.

• Designed a Python-based feature engineering service to generate behavioral risk signals (e.g., velocity checks, device fingerprinting),

with Redis caching for fast, real-time access.

• Deployed services using Docker and Kubernetes (AWS EKS) with Terraform for infrastructure and GitHub Actions CI/CD, ensuring

reliable releases with strong test coverage.

• Strengthened platform security with OAuth 2.0, JWT authentication, rate limiting, and data encryption, aligning with PCI-DSS and

GDPR requirements.

• Partnered with data science and product teams in an Agile setup to retrain and evaluate models using Snowflake data, improving

model performance (F1-score from 0.81 to 0.89) and enabling smooth, zero-downtime deployments.

• Engineered and scaled multiple cloud-native, high-availability applications supporting 5M+ active users, leveraging React, Next.js,

Node.js, and AWS, ensuring performance, resilience, and seamless user experience.

• Designed and developed modular, reusable frontend architecture using React Hooks, TypeScript, and component-driven design,

improving code maintainability and reducing feature delivery time.

• Built and integrated RESTful APIs and real-time WebSocket services, utilizing Redux for state management, enabling responsive and

scalable user interactions across distributed systems.

• Optimized backend performance by refining API queries, implementing caching strategies (Redis), and improving PostgreSQL query

efficiency, resulting in faster response times and reduced system load.

• Implemented server-side rendering (SSR) and dynamic imports in Next.js, along with WebP-based image optimization, significantly

improving page load speed and Core Web Vitals.

• Re-architected a third-party chat dependency into an in-house real-time messaging system using WebSockets and Redis Pub/Sub,

reducing operational costs by $40K+ annually while improving system control and latency.

• Contributed to the Pion WebRTC open-source ecosystem, enhancing real-time communication capabilities and collaborating with a

global developer community on performance improvements.

• Designed and deployed an event-driven notification system using AWS SQS and asynchronous processing, ensuring reliable message

delivery and system decoupling across services.

• Led the transition from a monolithic architecture to microservices-based design, improving scalability, fault isolation, and accelerating

release cycles through independent service deployments.

• Built and maintained a containerized CI/CD pipeline using Docker and AWS ECS, enabling multi-stage deployments, automated testing,

and high system availability in production environments.

• Worked on migrating on-premise production systems to IBM Cloud, handling environment setup, application deployment, and

validation; helped improve system stability and reduced infrastructure-related risks.

• Built and enhanced user-facing applications using React.js, focusing on reusable components, clean UI design, and smooth integration

with backend REST APIs.

• Contributed to backend development using Node.js and Express.js, supporting API development, data processing, and application

functionality across multiple modules.

• Improved database performance by refining data models, adding indexes, and optimizing complex SQL queries, which led to faster

response times and better application performance.

• Developed a MERN stack internal portal hosted on AWS (EC2, S3) to track migration progress and generate reports, reducing manual

effort by around 40%.

• Collaborated closely with cross-functional teams in an Agile environment, participating in sprint planning, code reviews, and

deployments to ensure timely and high-quality deliverables.