Working as part of the Logistics (Imports) team under the Stock & Pricing Technology domain, focusing on modernizing and securing logistics applications. I’ve worked on strengthening and modernizing key components of the Inbound Logistics System. I delivered a full RBAC security feature to manage user access more effectively and refactored legacy procedural code into a clean, object-oriented architecture for better scalability and maintainability. My work also involved setting up SQL Server databases, configuring IIS environments, and deploying internal documentation tools with secure Windows Authentication. I built RESTful API services using z/OS Connect to expose logistics cost data from mainframe systems to web applications, bridging modern and legacy infrastructure. Beyond development, I documented technical specifications, supported app releases, and led structured deployment processes with risk assessments and implementation plans to ensure reliable production rollouts.
Designed the foundational system architecture for an Agentic AI Medical Triage System from scratch. Translated clinical and stakeholder requirements into actionable technical specifications and documentation. Researched and evaluated various LLM, text-to-speech (TTS), and automatic speech recognition (ASR) models to determine the optimal AI stack for real-time medical triage. Developed a real-time conversational voice bot using Pipecat and Daily to perform symptom assessment, preliminary diagnosis, and triage for primary care (GP). Deployed the voice bot and backend services on Cerebrium AI's serverless infrastructure platform for scalable, low-maintenance operation.
Conducted comprehensive testing across development, pilot, and stable environments of the DoublOne web application, documenting and tracking bugs to ensure product quality. Introduced responsive web features and reusable components using React.js, and rebuilt DoublOne's website from scratch based on a new Figma design. Optimized the website codebase by applying SOLID principles, reducing code volume by around 25% (from 460 lines to 115 lines) to improve maintainability and efficiency.
Implemented RESTful API integrations to seamlessly connect frontend widgets with backend services. Developed real-time data ingesting applications using Firebase and the Flutter framework while following best practices such as SOLID, TDD, and clean code principles. Ensured designs were compliant with specifications and easily extendable by writing reusable, testable, and efficient Flutter code. Implemented robust state management and business logic for complex use cases using Provider, GetX, and Riverpod to deliver expressive and smooth user experiences.