Kavya Naidu

Deployed Java-based microservices on Apache Tomcat behind an API Gateway with authentication and rate limiting, handling 1.2M+ monthly API calls for payments, notifications, and session control. Integrated Kafka streaming with Python anomaly detection models (XGBoost, Isolation Forest) to scan transaction events in real time and automatically flag suspicious patterns, reducing fraud-related false negatives by 27%. Built Spring Boot REST APIs to serve CNN image classification outputs for automated product tagging. Developed secure Node.js middleware for role and access management in internal AI model dashboards, improving version traceability and audit compliance. Containerized Python scoring logic with Docker and orchestrated daily batch inference via Airflow, integrating outputs into a pricing engine for dynamic pricing across four customer cohorts. Wrapped LLM-driven summarization models into production-ready APIs with backend integration and failover support, cutting average custome

Built Java Spring Boot REST bridges to securely transmit inputs to GenAI services hosted on AWS SageMaker for features like document summarization and entity extraction within transactional workflows. Designed Java Kafka consumer services to process invoice creation/update events and trigger Python AI models for automated categorization, reducing manual classification workload by 40%. Implemented failover logic in Java microservices to detect AI API timeouts/failures, serve cached fallback outputs, and log detailed error trails for monitoring to maintain availability during model latency spikes. Scheduled daily batch jobs using Spring Scheduler to call LLM APIs for summarizing transaction descriptions, improving weekly report readability and accuracy. Integrated AI-generated invoice suggestions into an existing Java-based validation pipeline with tax rule compliance checks before publishing to user dashboards within the QuickBooks ecosystem.

Developed RESTful APIs using Flask and FastAPI for authentication, product inventory, and MIS reporting, reducing API response time by 30% and enabling smooth frontend integration. Built responsive role-based dashboards and admin panels using HTML5, CSS3, JavaScript, and Jinja2 templates. Wrote optimized SQL queries and stored procedures in PostgreSQL/MySQL to support complex joins, filtered views, and real-time reporting for operations and finance. Automated recurring tasks (invoice generation, daily backups, system alerts) using Celery + Redis, reducing manual ops load and ensuring timely stakeholder notifications. Modeled data using SQLAlchemy ORM, implemented migration scripts, and maintained schema version control across dev, staging, and production. Containerized the application stack with Docker and used Docker Compose during testing/UAT. Implemented multi-level role-based access control across API and UI layers with strict database read/write permissions for sales, admin, and f