cv

IT Support Officer

World Health Organization

Administered DHIS2 and OpenMRS infrastructure supporting the national HIV/AIDS eHealth program for 1,000+ users, in collaboration with UNAIDS, ICAP, HISP India, and CHAI.

• Maintained 99.99% uptime on healthcare data servers; managed user accounts, metadata, and national health data reporting workflows for on-premises and cloud-based systems
• Designed and maintained data workflows from field data collection through national health reporting, ensuring data accuracy, traceability, and compliance with UN governance standards
• Led user onboarding, IT asset management, and cybersecurity initiatives; collaborated with UN agencies and national health authorities on data governance and compliance standards
• Stack: PostgreSQL · Java Tomcat · NGINX · Linux · DHIS2 · OpenMRS

Mobile Application cum IT Technician

CESVI Myanmar

Built and maintained end-to-end data pipelines for mobile ICT-enabled extension services reaching 130 villages and 26,234 households across Myanmar's Dry Zone.

• Co-developed GIS dashboards and data workflows from KoBoToolbox field collection to donor reporting for 6 agriculture and rural development projects funded by UNDP, FAO, SDC, AICS, and CESVI
• Digitized agro-advisories into mobile formats, enabling data-driven decision-making for rural farmers
• Trained 2,000+ farmers in mobile application use and digital literacy
• Stack: KoBoToolbox · QGIS · Looker Studio · Esper · Power BI

IT Service Engineer

Huawei Technologies Co., Ltd.

Delivered and commissioned 7 nationwide carrier and enterprise IT infrastructure projects for MPT, Ooredoo, Telenor, and CB Bank.

• Maintained MPT Private Cloud operations (ECS, EVS, VPC, OceanStor); configured Eudemon 8000E firewalls, VPNs, DNAT/SNAT, and SAN zoning
• Developed technical documentation (LLDs, HLDs, SOPs, MOPs) and trained client staff; collaborated with Huawei R&D and GTAC for SLA compliance
• Stack: FusionCloud · FusionCube · NetEco · OpenStack · Huawei Cloud

Junior Network Engineer

ISGM (ICT Star Group Myanmar) Co., Ltd.

Supported NEC Japan engineers in infrastructure audits and optical network expansion across metro and national backbone networks.

• Managed blade servers, virtualization platforms, storage systems, and Fortigate firewall configuration
• Implemented wireless (Unifi AP), monitoring (Zabbix), and authentication (NEC SFA) systems

Junior Network Engineer (Internship)

Netpro Myanmar Co., Ltd.

Assisted senior engineers in deploying and configuring enterprise network infrastructure for internal systems and client projects.

• Gained hands-on experience with VoIP (Yeastar), CCTV (Dahua, HIKVISION), GPS tracking (Teltonika), and network equipment (MikroTik, Ubiquiti)

Master of Science

Batangas State University

Data Science

• IIE Myanmar Student Emergency Support (MSES) Fund Grantee — Institute of International Education
• Thesis: Development of Hybrid Lightweight CNN Integrating SqueezeNet and CBAM for Edge Deployment Using Multi-Data Sources
• Fundamentals of Data Science
• Big Data and Cloud Computing
• Machine Learning and Neural Networks
• Time-Series Analysis
• Mathematical and Computational Theories
• Seminars in Data Science

Bachelor of Computer Technology

University of Computer Studies, Mandalay

Computer Communication and Networks

• Data Mining
• Programming (C, C++, Java)
• Data Structures & Algorithms
• Systems Analysis & Design
• DBMS
• Networking & Security
• Advanced Networks
• Linear Algebra
• Statistics
• Embedded Systems & HCI
• Project Management

Wikipedia Web Traffic Prediction via Time Series Modeling

Journal (Under Revision)

Comparative study of 7 time-series forecasting models — Moving Average, Exponential Smoothing, ARMA, ARIMA, Auto-ARIMA, Prophet, and LSTM — on a 145,000-article Wikipedia Kaggle dataset. Co-authored with Prof. Sandhya Avasthi, ABES Engineering College, India.

Shwe Taungthuu — Myanmar Agricultural Advisory Agent

AI-powered bilingual agricultural advisory app for Myanmar farmers. Built with Google Gemini 2.5 Flash via Vertex AI; Python + FastAPI backend deployed on Google Cloud Run. Presented at Google Developers Community Build with AI MMDT 2026.

• Burmese voice and text input with structured advisory responses in real time
• Voice History feature for replay of original queries — designed for low-literacy, field-usability contexts
• Keyless cloud authentication via Application Default Credentials (ADC)

Chickpea Fusarium Wilt Detector

MSc thesis research — hybrid lightweight CNN for on-device plant disease detection. Fully offline bilingual Android app (English + Myanmar Unicode) for smallholder farmers. Full methodology reserved for upcoming publications.

• 82.09% cross-domain accuracy at 0.870 MB; validated on 115 real Android devices via Google AI Edge Portal
• Worst-case inference: 73.7 ms on low-tier CPU-only hardware (Android 9.0+, 2 GB RAM compatible)
• Personally collected field dataset from Myanmar farming communities with expert agronomist annotation

Bean Disease Classification via Stepwise CNN Fine-Tuning

Benchmarked 5 pretrained CNN architectures; designed a stepwise fine-tuning strategy achieving 98.86% test accuracy on a combined bean leaf disease dataset.

• Identified ResNet50 as optimal backbone through systematic comparative evaluation
• Stepwise layer unfreezing to adapt ImageNet representations to the plant disease domain

Wikipedia Web Traffic Prediction via Time Series Modeling

Research paper under revision comparing 7 time-series forecasting models on a 145,000-article Wikipedia Kaggle dataset.

• Evaluated Moving Average, Exponential Smoothing, ARMA, ARIMA, Auto-ARIMA, Prophet, and LSTM using RMSE
• Identified Moving Average as the best-performing model, showing simpler statistical approaches outperformed LSTM on this task
• Co-authored with Prof. Sandhya Avasthi, ABES Engineering College, India

MMDT Intro to Deep Learning Course Projects

Four progressive projects covering tabular regression, NLP classification, CNN benchmarking, and transformer-based geocoding. Completed with A+ grade.

• Fine-tuned BERT-multilingual, DistilBERT-multilingual, and XLM-RoBERTa on 610,509 Myanmar addresses for coordinate regression
• Deployed A2C live demo on Hugging Face Spaces with Streamlit, Plotly maps, and kilometer-level error comparison
• Built HDB resale price prediction with MLP regression achieving R2=0.92 and RMSE=SGD 59,051

ML Algorithms from Scratch

5 classical ML algorithms implemented from first principles using NumPy — no sklearn for core logic. MSDS coursework covering formula derivation, manual computation, and validation on real datasets.

• Linear Regression, Logistic Regression, Decision Tree, KNN, Naive Bayes — each with step-by-step written explanation and figures
• Naive Bayes 90.00% (Iris), KNN Best k=5 at 72.73% (Pima Indians), Decision Tree Train 78.57% / Test 71.05% (Iris)

Loan Approval Prediction using Machine Learning

MSDS coursework project building machine learning workflows to improve loan approval accuracy and reduce manual decision bias.

• Compared Logistic Regression, Random Forest, SVM, Gradient Boosting, and Decision Tree with 10-fold cross-validation
• Best model: Logistic Regression with Accuracy 0.811 and AUC 0.757
• Identified credit history and applicant income as key drivers, with limited impact from demographic features

Python-Powered Data Cleaning with Pandas, NumPy and Dask

Scalable data cleaning coursework project transforming raw, inconsistent datasets into analysis-ready data for big data systems.

• Implemented median imputation, duplicate removal, and IQR-based outlier detection
• Benchmarked Pandas vs Dask on large datasets including a 2.16 GB Toxic Release Inventory dataset
• Validated downstream improvements with regression metrics and exploratory visualizations