Steve Ngai

Arvato Systems Malaysia is a part of Arvato Systems, providing IT solutions and services to support digital transformation, focusing on industries like media, retail, utilities, and healthcare.

Key Responsibilities

• Contribute to the full software development lifecycle by writing clean, maintainable code, creating automated tests, debugging issues, and optimizing performance. Collaborate with cross-functional teams in an Agile Scrum environment, participating in sprint ceremonies and maintaining comprehensive technical documentation such as database schemas and flow diagrams to ensure quality, clarity, and continuous improvement.

Implemented Solutions

• POC Implementations (2024): Led two POCs:
  • Azure Monitor / Application Insights for enhanced observability.
  • Google Ad Manager API Dev Tool to accelerate development and testing for Digital Sales.
• Completed POC (2025): Completed Hangfire-based centralized Job Scheduling in .NET 10.
• Ongoing POC (2025): Developing AI Model Monitoring & Automated Retraining for Google Ad Click Prediction.
• Repository Optimization: Reduced repository size by 24% by cleaning up the package folder.
• Critical Blocker Resolution: Enabled line item status transitions, eliminating the need for costly external consultancy.
• CI/CD Optimization: Implemented a file-based (auto) test-skipping mechanism and added manual test-skipping via [skip-test] in commit messages, reducing CI pipeline time by 66%.
• CI/CD Migration: Migrated build pipelines from Azure DevOps (Windows) to GitHub Actions (Ubuntu), refactoring YAML workflows and build scripts to reduce CI/CD runtime by 50%.
• Security: Proposed and implemented a major code refactoring of the application's security layer by centralizing access control into a unified security service, replacing manual validations, resolving critical IDOR vulnerabilities, and improving maintainability.
• Security: Implemented a secure logging feature in product framework by introducing masking policies (e.g., email and GUID), enhancing logger configuration, and ensuring compliance with security/privacy requirements while maintaining debugging effectiveness.

Extra Initiative

• AI Hackathon Participation: Competed in Arvato Systems Malaysia: AI Hackathon 2024, collaborating on an AI POC for support ticket management using Python, Google Colab, and JIRA API integration.
• Team Engagement: Volunteered to organize a team-building event for the BMS team during counterpart visits to Kuala Lumpur.
• Knowledge Sharing & Tech Talks:
  • Regularly shared on weekly GitHub Copilot knowledge sharing sessions for a team of 16, to encourage AI-assisted development adoption.
  • Conducted three knowledge-sharing sessions on IDOR security, ReSharper and modern C# development practices, and modern documentation techniques using Mermaid.
  • Regularly initiated cross-team discussions to improve code quality, introducing:
    • Strategy & Factory Method patterns for better maintainability.
    • Pure unit testing as a new approach.
• AI-Driven Productivity: Leveraged GitHub Copilot to accelerate feature implementation and unit test creation, code refactoring, documentation generation, and debugging assistance.
• API Documentation Improvement: Proactively identified critical API documentation gaps, researched and proposed a modern, centralized web-based solution, gaining team lead recognition for its long-term strategic value despite current implementation constraints due to team restructuring.
• API Documentation Improvement: Enriched Swagger UI with auto-generated validation and examples, streamlining frontend development and minimizing clarification needs.

Accomplishments

• Performance Recognition: Successfully passed 6-month probation with strong results and achieved a good performance review for 2024.

Technology Stack

• Languages & Frameworks: C#, .NET Framework 4.8, .NET 8, Web API
• Backend Tools: Swagger (Swashbuckle), Hangfire, Serilog, Seq
• Databases: SQL Server, Entity Framework Core 8
• Cloud & DevOps: GitHub, Azure DevOps, Azure Function, Azure Service Bus
• Monitoring & Testing: Azure Application Insights, NUnit, MsTest (Unit Test, Integration Test)
• Frontend: JavaScript
• Design Patterns & Principles: Monolith, Microservices, Factory Pattern, Strategy Pattern, Query Object Pattern, Repository Pattern, DRY, YAGNI, SOC, OOP, Dependency Injection (DI), MVC
• Others: Scrum, Agile, Confluence, MermaidJS, SonarQube / SonarCloud, Google Colab, Python

• Oct 2025: Implemented an end-to-end MLOps pipeline in Azure Machine Learning and Azure ML Studio using the Azure ML SDK and MLflow; automated data preparation, model training, and model registration. Established experiment tracking, environment versioning, and compute orchestration for reproducible model development and retraining. | Code.
• May 2025: Developed a configurable RAG pipeline in Python for HR policy question answering, utilizing Sentence Transformers for embeddings, ChromaDB for vector storage, and the LM Studio API. Implemented dynamic model selection for both embedding generation and answer synthesis, enabling seamless switching between models such as Zephyr and Llama 3.2, enhancing flexibility and adaptability. Employed Langchain's RecursiveCharacterTextSplitter with prioritized separators for effective text chunking and similarity search for relevant context retrieval | Code.
• May 2025: Configured and deployed LM Studio with Ollama 3.2 API server locally, serving the llama-3.2-3b-instruct model; developed a Python client to connect via HTTP API and retrieve AI-generated responses | Demo | Code.
• Jan 2025: Developed a POC using Python and Retrieval-Augmented Generation (RAG) with a Pinecone-based assistant integrated with LLM GPT-4. Enabled natural language querying of PDF documents, including ebooks and HR policies, with responses strictly derived from uploaded content | Demo.
• Jan 2025: Developed a POC using the BART transformer model with Hugging Face to summarize PDF text. Integrated PyMuPDF for extraction, Streamlit for an interactive UI, and optimized text processing to address encoding challenges | Demo.
• Oct 2024: Developed a POC which highlights the asynchronous nature of order handling and the integration of Azure Service Bus for messaging using Azure Function App .
• Jun-July 2024: Deployed an eShop using .NET Aspire to both local and Azure environments. Utilized Azure Developer CLI (azd) to provision 15 Container Apps within 11 minutes .
• May 2024: Deployed a minimal starter pipeline in Azure DevOps to build, test, and deploy an ASP.NET Core project, utilizing stages for artifact download and deployment to an Azure Web App.
• July-Aug 2023: Built a Dagster ETL data pipeline in Python, scheduled to run hourly. Fetched data from Hacker News API and generated CSV, and read data from SQL Server.
• June 2023: Configured eShopOnContainers with Azure Kubernetes Service (AKS) for CI/CD using GitLab Actions. Integrated Azure Active Directory for authentication.
• June 2023: Deployed .NET 7 eShopOnContainers to Local Kubernetes using Docker Desktop and Helm.
• May 2023: Built a mock eCommerce microservices solution using GraphQL, Azure Functions, and Azure Service Bus.
• May 2023: Developed Azure Functions demo projects (Blazor Server + Web API). Integrated Azure AppInsights for monitoring.
• Mar 2023: Optimized online resume for 100/100 GTmetrix performance and 90/100 in Pingdom.
• Mar 2023: Deployed a Jekyll-based website with GitHub Actions CI/CD.
• Feb 2023: Built a mock Ethereum blockchain Web API using .NET 7 and EF Core in 3-4 days.
• Nov 2022: Automated code reviews in Azure DevOps CI pipeline with SonarQube and SonarCloud.
• Aug 2022: Deployed .NET 6 Web API to Azure using GitHub Actions (CI/CD).
• July 2022: Implemented and monitored automated tests in Azure DevOps CI/CD pipeline.
• May 2022: Deployed Task Tracker API in .NET 6 and refactored Task Tracker React app for API consumption.
• Oct 2021: Integrated SpecFlow + Living Doc in .NET 6.
• July 2021: Implemented SonarCloud (Code Analysis) with GitHub Actions .
• Jun 2021: Migrated online resume to Pug + SCSS, reducing file size by 40% and improving maintainability.
• Apr 2021: Parsed 1,000 complex XML files into Elasticsearch using NEST C# in 15 minutes with advanced search options.
• Feb 2021: Completed take-home technical assignment in 24 hours (given 48 hours). Built a .NET Core Web API service to upload, parse, and query transactions from XML & CSV.
• Jan 2021: Parsed 800 XML files (500MB) into MongoDB using Node.js in 1 minute . Parsed 1,800 XML files (4.7GB) into MongoDB using C# in 9 minutes. Ingested XML files into Elasticsearch using Node.js & Python.
• Nov-Dec 2020: Built a JWT-based CRUD web app (Back-end: .NET Core 3.1, Front-end: Vanilla JS).

Quantum Corporation sells scale-out storage, archive, and data protection solutions for capturing, sharing, and preserving digital assets in physical and virtual environments.

Key Responsibilities

• Coding & Development: Write clean, efficient, and maintainable code.
• Testing & Debugging: Write automated unit and integration tests, and fix bugs to ensure software quality.
• Collaboration: Work with cross-functional teams, including designers, product owners, and QA engineers.
• Documentation: Create technical documentation like database schema diagrams, sequence diagrams, and flowcharts for future reference.
• Performance Optimization: Improve the efficiency of applications.
• Automation: Automate manual processes by implementing generic background microservices.
• Data Migration: Perform ETL data migration to ensure seamless data transfer across systems.

Implemented Solutions

• Automated Fulfillment: Developed a background microservice to reduce manual processing.
• IT Asset Management: Built a CMDB system for better IT asset tracking.
• Automated Testing: Introduced and implemented automated unit and integration tests.
• ETL Migration: Developed a .NET 6 ETL service to migrate data from MS Dataverse to SQL Server, cutting Azure costs by ~$3,000/month.
• Performance Optimization: Upgraded ETL migration (40M records) from .NET 6 to .NET 8, achieving:
  • 16% faster bulk operations (7.3M records)
  • 30% faster smaller transactions (630K records)
• Process Automation: Built a static HTML email utility and Power Automate workflows to streamline internal tasks.

Extra Initiative

• Tech Exploration: Researched SonarQube (code quality) and Dagster (data orchestration) for future improvements.

Accomplishments

• Cost Reduction: Cut Azure costs by ~$3,000/month through ETL optimization.
• Performance Gains: Boosted ETL efficiency by up to 30% in .NET 8 with no added infrastructure cost.

Technology Stack

• Languages & Frameworks: C#, .NET 6, ASP.NET Razor Pages, MVVM
• Backend Tools: Hangfire, Entity Framework Core
• Databases: SQL Server, MS Dataverse
• Monitoring & CI/CD: Azure App Insights, Health Checks, GitHub Actions

AFFIN Hwang Investment Bank Berhad provides financial services. The Company offers stock brokerage, asset management, treasury bonds, currency exchange, financial advisory, and investment banking services.

Key Responsibilities

• Coding & Development: Write clean, efficient, and maintainable code.
• Testing & Debugging: Write automated unit tests, automated integration tests, and fix bugs to ensure software quality.
• Collaboration: Work with solution architect, CTO, front-end tech lead, back-end tech leads, business analyst, UI/UX designer, data engineer, and end-users.
• Documentation: Create technical documentation like database schema diagram, sequence diagram, and flowchart for future reference.
• Performance Optimization: Improve the efficiency of applications or services.
• Interviewing: Conduct interviews for software engineers.

Implemented Solutions

• Microservice Development: Replaced a monolithic legacy system (.NET Framework 4.x) with a microservice-based automated data ingestion solution using .NET Core 3.1, improving scalability and efficiency.
• Performance Optimization: Enhanced the performance of eInvest Go and the Retail Online Platform by optimizing both front-end and back-end services for better responsiveness and reliability.

Extra Initiative

• Process Improvement: Initiated and developed productivity tracking reports in Jira to monitor team performance.
• Versatility: Transitioned from backend to full-stack development by successfully taking over frontend responsibilities in Vue.js, despite the challenging learning curve.

Accomplishments

• Consistently High Performance: Consistently exceeded expectations in performance appraisals from Aug 2020 to Dec 2022.
• Business Impact: Improved system reliability through REST API development and automated testing, leading to a 20% increase in online trading account openings (eInvest GO).

Technology Stack

• Backend (70%): C# Web Api, .NET Framework 4.x, .NET 2.2, .NET Core 3.1 , EF Core 3 , NUnit, Xunit , Specflow and Living Doc.
• Frontend (30%): VueJs (SFC), Vuex, Javascript, Jquery, Bootstrap, Pug, Scss, NPM, Gulp, Razor Pages.
• Others: Git, Atlassian Jira, Confluence, Azure DevOps, IIS, Figma, Elastic Search, Kibana, NEST, Google Cloud Storage (Google Cloud Platform), mermaidJS, MarkMap.

Eduvo is a vocational school for IT and Business courses.

Key Responsibilities

• Design and develop training programme (Prof. Diploma in Software Engineering).
• Conduct classroom training for software engineering courses: Object-Oriented Programming, Database Design, Web Programming, System Analysis and Design, Data Structure and Algorithm, and Software Architecture.
• Design, develop, and maintain School Management System.
• Lead Education & Operation division.

Implemented Solutions

• Developed a School Management System from scratch using C#, including an automated timetable generation feature that schedules all courses for the entire year and assigns each course to a lecturer.

Accomplishments

• I pioneered and was the first to design and develop the Software Engineering Prof. Diploma programme.
• Software engineering student enrollment grew by 300%, with graduates now employed as full-time programmers in software houses.

Technology Stack

• Backend (70%): C# MVC, Entity-Framework (Code-First), SQL, LINQ
• Frontend (30%): Javascript, CSS, HTML.

GHL is a leading payment service provider and one of the top merchant acquirers in the ASEAN region. We provide world-class payment services and solutions. My job scope are:

• Reported to CTO.
• Created complex monthly IT Resource Utilization report for CTO.
• Managed Product Management department.
• Ensured product gets certified, especially Terminal Quality Management (TQM).
• Handled software and product documentation, i.e user guide, sequence / process flow diagram and matrix diagram.
• Ensured proper source control and software repositories backup and replicated for all the regions (Malaysia, Thailand, Philipinnes and Wuhan).
• Planned and coordinated training for Level 1 and Level 2 Support.
• Maintained maintenance agreement.