Ze Wei NG

• Solutioned and deployed an infrastructure architecture design pattern on private and public cloud
• Proposed 3 solutions to improving the Virtual Desktop Infrastructure experience locally, backed by extensive data analytics and communication with global stakeholders
• Increased efficiency for business stakeholders through close collaboration and agile development deployment of VBA scripts for process automation

• Led a 6-figure E2E product lifecycle, delivering the product on time and within budget
• Coordinated effective communication with the client, 2 external vendors, and a dozen end-users, overseeing successful deployments in Development, UAT and Production environments
• Solutioned and deployed an infrastructure architecture design pattern on private and public cloud
• Wrote and deployed Unix bash scripts for encryption and file transfer automation; developed scalable applications using Pega low code and Node.js

• Identified and presented 3 solutions to reduce incidents for a Treasury application by over 50% through data-driven visualization and analysis of incident logs
• Achieved business efficiency through automating the identification of box jobs within Autosys which fail to meet time and status criterion through Unix bash, Python, and PowerShell scripting and scheduled hourly email alerting in HTML

• Achieved Spoken Language Identification of South East Asian languages through Python (PyTorch, MATLAB Engine API) and Unix bash scripting
• Designed and trained a neural network which achieved f1 scores of 95.3% for in domain testing and 72.7% for out of domain testing
• Curated and processed over 500 hours of open source South East Asian speech data, and published train, test, and validation datasets on Kaggle

NextJS webapp integrated with a Python backend, integrating CV capabilities for audio/music score alignment using Conditional YOLO, and audio-audio alignment through Dynamic Time Warping.

Achieved sub-2 metre localization accuracy in indoor environments using Machine Learning algorithms and Deep Learning models, ESP32 microprocessors, C++ programming, and UDP networking for real-time front-end localization and visualization.

Created and demonstrated a POC SCM tool through Hyperledger Fabric DLT backend networking, JavaScript, Docker, and POST APIs.