My Portfolio
Projects
- All
- Python
- AI Automation
PowerBuilder to .NET C# API
+PowerBuilder to .NET C# API Migration
Description: A pioneering project migrating legacy PowerBuilder systems with complex DataWindow objects to modern .NET C# APIs, utilizing Agentic AI Pipelines for autonomous, intelligent refactoring.
My Role: Backend Developer — Designed and implemented the Agentic AI Pipeline, managing autonomous workflows, custom AI skills, and ensuring 24/7 production stability.
Tech Stack: Python, AI Agents, .NET C#, Docker, Generative AI.
Key Features:
- Autonomous Orchestration: AI Agent-driven end-to-end conversion workflow.
- Custom AI Workflows: Specialized agents with skills, rules, and guardrails for accurate refactoring.
- Hybrid Logic Engine: Integrates generative AI with Python scripts for deterministic mapping.
- Resource Management: Intelligent monitoring of token availability and build states.
Challenges:
- Maintaining 100% accuracy in complex business logic translation.
- Handling massive codebases with minimal human intervention.
Solutions:
- Implemented AI guardrails to minimize errors and ensure best practices.
- Built high-availability backend with operational excellence.
Results:
- Dramatically accelerated migration timeline.
- Achieved consistency and scalability in legacy modernization.
TexManage
+TexManage - Textile Business Management
Description: An application for managing customers, orders and inventory for Vietnamese textile businesses.
My Role: Backend Developer - Developed the Django Apps and managed the database.
Tech Stack: Python, Django, SQLite, ChartsJS.
Key Features:
- Automated order and detailed inventory management.
- RESTful API integration with third-party applications.
- Dashboard reporting with ChartJS for business overview.
Challenges:
- Handling large datasets and optimizing API performance.
- Looping nested URLs and design database due to many types of textile and
colors.
Solutions:
- Used caching and optimized SQL queries.
Results:
- Reduced order processing time by 30%.
MediFlow
+MediFlow - Medical Records Automation
Description: Developed a Python automation tool to download and compare current data with historical records, identifying potential misspellings or incorrect dates of birth using fuzzy matching. Integrated Microsoft Azure Graph API to notify clients of results via email. Automated daily execution with Task Scheduler and enabled manual triggers through email keywords.
My Role: Backend Developer — Designed and developed Python automation solutions from the ground up, building backend systems to enhance efficiency and accuracy.
Tech Stack: Python, Selenium, Pandas, Fuzzy, Graph API, Task Scheduler, XLSXWriter
Key Features:
1. Data Extraction and Comparison
- Automated the download of current date's data.
- Compared new data with historical records to identify discrepancies.
2. Fuzzy Matching for Data Accuracy
- Implemented fuzzy matching algorithms to detect misspelled names or incorrect
dates of birth.
3. Automated Notifications
- Integrated Microsoft Azure Graph API to send results to clients via
email.
4. Scheduled and Triggered Execution
- Configured Task Scheduler for daily automation runs.
- Enabled manual triggers via specific keywords in received emails using
Python.
5. Error Handling and Logging
- Implemented error handling and logging mechanisms for seamless operation and
troubleshooting.
Challenges:
- Slow performance of fuzzy matching algorithms.
- Risk of false positives when flagging potential errors.
Solutions:
- Optimized fuzzy matching: Limited it to a subset of data with
an 80% match on dates of birth, focusing checks only on potential misspelled
names.
- Used simple comparisons to detect incorrect dates of birth with matching
names.
- Green List Implementation: Developed a whitelist feature
allowing clients to override and confirm accurate patient data.
Results:
- Reduced fuzzy processing time by 15 times — from 30 minutes
to 2 minutes.
AxcessDash
+AxcessDash - Django Dashboard for CCH Axcess
Description: Developed a Django application that integrates with the CCH Axcess API, providing authentication via username and password. Implemented token-based authentication to perform tasks such as retrieving client lists, downloading documents, and checking eFile tax form status. Replaced the outdated CCH Axcess desktop installation and UI with a modern, web-based interface for improved user experience and accessibility.
My Role: Backend Developer — Built the project from the ground up, designing and implementing backend systems to integrate with CCH Axcess API seamlessly. Focused on enhancing user experience with a clean, responsive web UI to replace the legacy desktop interface.
Tech Stack: Python, Django, SQLite, CCH Axcess API, Bootstrap
Key Features:
1. User Authentication
- Implemented login using username and password with token-based
authentication.
2. Document Management
- Enabled downloading of documents directly from CCH Axcess.
3. Client Management
- Retrieved and displayed client lists efficiently.
4. eFile Status Tracking
- Integrated features to check and display the status of eFile tax
forms.
5. Modern Web UI
- Designed a clean, user-friendly web interface to replace the outdated CCH
Axcess desktop UI, enhancing accessibility and user experience.
Challenges:
- Authentication Complexity: Dealing with complex token-based
authentication.
- API Documentation Issues: Poor and unclear API documentation
with missing information.
- Legacy UI Limitations: Overcoming the limitations of CCH
Axcess's outdated desktop interface.
Solutions:
- Authentication Handling: Built a robust token management
system to streamline authentication processes.
- API Integration: Created custom API wrappers and
error-handling mechanisms to manage unclear and missing information
effectively.
- UI Modernization: Developed a responsive web-based UI using
Bootstrap, providing better access and usability compared to the legacy desktop
version.
Results:
- Improved API Integration: Enhanced API call efficiency and
reliability.
- Streamlined User Experience: Simplified access to documents,
client lists, and eFile status.
- Modernized Interface: Successfully replaced outdated CCH
Axcess desktop installation with a user-friendly web app, improving
accessibility and efficiency.
SnapChart
+SnapChart - Live Chart Django Application
Description: Developed a Django application capable of updating real-time data and capturing screenshots every hour, providing users with timely visual insights.
My Role: Backend Developer — Built the screenshot capture solution from the ground up and integrated Chart.js for dynamic data visualization.
Tech Stack: Python, Django, SQLite, Chart.js
Key Features:
1. Real-Time Data Updates
- Enabled automatic updates for charts with live data.
2. Scheduled Screenshot Capture
- Implemented hourly screenshot functionality to archive data visuals.
3. Interactive Charts
- Integrated Chart.js for responsive and interactive data visualization.
4. User-Friendly Dashboard
- Designed a clean and accessible dashboard for managing and viewing charts.
Challenges:
- Efficient Screenshot Capture: Ensuring low-latency and
resource-efficient screenshot captures.
- Data Syncing: Maintaining consistency between real-time data
updates and chart displays.
Solutions:
- Optimized Capture Method: Utilized asynchronous tasks to
capture screenshots without interrupting data updates.
- Data Buffering: Implemented data buffering to sync real-time
data with charts smoothly.
Results:
- Improved Performance: Reduced screenshot capture time by
optimizing tasks.
- Enhanced User Experience: Delivered a seamless and
interactive charting experience with real-time updates.
ExcelFlow BI
+ExcelFlow BI — Data Processing for Power BI
Description: Developed a Python automation tool to streamline the process of downloading, renaming, and organizing Excel files based on current dates, facilities, and practices. Prepared the data for seamless import into Power BI dashboards.
My Role: Backend Developer — Built the download functionality from the ground up, ensuring efficient file management and preparation for Power BI integration.
Tech Stack: Python, Selenium, Power BI, Task Scheduler
Key Features:
1. Automated File Download
- Implemented a Python script to download Excel files automatically based on
schedule.
2. Dynamic File Management
- Renamed and moved files into appropriate folders based on date and
facility/practice.
3. Data Preparation for Power BI
- Pre-processed and cleaned Excel files to ensure smooth import into Power BI
dashboards.
4. Scheduling and Task Automation
- Utilized task scheduling to automate file downloads and preparation at regular
intervals.
Challenges:
- Inconsistent File Formats: Handling variations in Excel file
formats from different sources.
- Efficient Scheduling: Ensuring the automation runs smoothly
without manual intervention.
Solutions:
- Format Handling: Developed scripts to standardize and clean
data before processing.
- Optimized Scheduling: Configured task scheduler to automate
downloads and data preparation reliably.
Results:
- Improved Efficiency: Reduced manual workload for file
management and preparation.
- Seamless Integration: Enabled smooth and error-free data
imports into Power BI dashboards.