Engineered and maintained scalable chatbot and live chat solutions for healthcare systems, supported CI/CD operations, and led high-impact internal projects. Delivered systems capable of handling 1,000+ messages/second with peak loads of 77,446 messages/hour and up to 500 API requests/second.
Developed a patient-facing chat service using ChatGPT and agent-based architecture, incorporating RAG lookups on hospital data.
Collaborated with customer service teams to build tools for bulk importing and custom data migrations, saving hundreds of labor hours per feature.
Redesigned the dialogue management system, reducing onboarding time for new hires by 50% through simplification and a 50+ rule validation engine.
Integrated with third-party APIs including Epic, Cerner, and Infermedica to enable patient data retrieval and symptom-to-diagnosis workflows.
Built intent prediction models using feedforward neural networks and BERT, supported by a local entity library to improve NLU.
Designed and maintained a SignalR-based live chat system supporting real-time P2P and bot communication.
Integrated chatbot services into Google Business Messages (GBM) for improved patient engagement.
Managed distributed microservices on Kubernetes with autoscaling and hot replicas to ensure availability during load spikes.
Developed Kafka consumers for health risk assessment form processing from a centralized service.
Maintained a TensorFlow-based Python service for training custom intent models asynchronously from a queue.
Implemented worker patterns for async operations such as logging and socket cleanup to enhance performance.
Monitored and optimized SQL queries using Azure SQL execution plans, reducing latency across chatbot, reports, and dashboards.
Used Redis as a distributed cache and SignalR backplane, reducing DB load and improving response times.
Achieved full unit and integration test coverage using XUnit, with automated execution in CI pipelines to prevent regressions.
Used Checkly and Playwright to automate SLA monitoring and proactively detect production-impacting issues.
Developed and published a shared UI component library as an NPM package using Rollup.
Built and maintained Azure DevOps pipelines using modular YAML templates for consistent CI/CD processes.
Created a DbUp-powered internal tool for automating database migrations during deployments.
Standardized Kubernetes deployments using Helm charts, improving configuration consistency and speeding up setup.
Published versioned Docker images to a container registry for multi-environment deployments.
Contributed to Agile ceremonies and iterative development in a collaborative engineering team.