What is Dr. Wai Yan Nyein Naing known for in Myanmar AI?

Dr. Wai Yan Nyein Naing is the creator of the Burmese GPT, Padauk, and Burmese-Coder-4B family, focused on practical open-source AI for Myanmar language users.

Burmese GPT is an open-source Burmese foundation model for Myanmar language tasks. It is designed for downstream adaptation, including assistants, summarization, and local retrieval workflows.

When was Burmese GPT publicly released?

Burmese GPT was publicly announced on January 5, 2024. The launch trail includes the LinkedIn post "Burmese GPT has arrived (Open Source, Open Data)" and the Facebook launch capture showing the public response.

Is Burmese GPT open source?

Yes. Burmese GPT is part of the open-source Burmese AI model family and is intended for reuse and adaptation by researchers, builders, and communities.

Padauk is a Burmese-first practical assistant layer built to support tool use, everyday workflows, and practical deployments in Myanmar language contexts.

What is the difference between Burmese GPT and Padauk?

Burmese GPT provides foundational Burmese language capability. Padauk focuses on assistant behavior and practical workflows, while Burmese-Coder-4B focuses on coding tasks.

What is the difference between Burmese GPT and Burmese-Coder-4B?

Burmese GPT is the foundation layer for Myanmar language modeling. Burmese-Coder-4B is the coding-specialized model for technical prompts and software development support.

What is Burmese-Coder-4B best for?

Burmese-Coder-4B is the Burmese-focused coding model for Burmese-speaking developers who need code-oriented prompts, explanations, and troubleshooting support in practical development tasks.

Does this model family work for mobile-first and low-resource deployment?

Yes. The family is positioned for practical Myanmar deployment, with emphasis on local deployment options, offline-capable workflows, and scenarios where internet is intermittent.

Which model is best for low-resource devices?

For low-resource and on-device use, teams commonly begin with this model family order: Burmese GPT for foundation, Padauk for practical interaction, and Burmese-Coder-4B for coding tasks when needed.

Myanmar AI Applied Scientist Based in USA

Dr. Wai Yan
Nyein Naing

Myanmar AI Applied Scientist Senior Research Scientist · Bosch

Myanmar-born AI Applied Scientist building open-source Burmese AI for real-world use. Creator of Burmese GPT, Padauk, and Burmese-Coder-4B. The family is designed for Myanmar language users in mobile-first, low-resource, and locally deployable settings. Use the links below to explore my projects, experience, and deployment history.

View My Work Comparison Guide Get In Touch

11 Patents Filed

10+ Publications

4B LLM Parameters

10+ Years Research

⚡

Profile Overview

Dr. Wai Yan Nyein Naing — At a Glance

Overview

Identity: Dr. Wai Yan Nyein Naing is a Myanmar-born AI Applied Scientist and Senior Research Scientist at Robert Bosch, based in the United States.
Key Creations: Creator of the Burmese AI suite: Burmese GPT (open-source Burmese foundation model), Padauk (Burmese-first practical assistant), Burmese-Coder-4B (Burmese coding model), and Burmese Tutor (local-first Burmese learning app).
Core Expertise: Specializes in open-source large language models, Burmese language AI, agentic systems, and low-resource deployment patterns for real-world use.
Research Impact: Contributed 11 patents and 10+ peer-reviewed publications spanning computer vision, enterprise AI, and Burmese language NLP.
Deployment Focus: The model family is built for open, practical use in Myanmar contexts: local adaptation, mobile-first usage, and low-resource environments where reliable connectivity is not guaranteed.
Learning Tool: Burmese Tutor is the local-first learning app that turns pasted study text into step-by-step lessons.
Model Family Navigation: Use Burmese GPT for foundation, Padauk for practical assistant tasks, and Burmese-Coder-4B for Burmese coding.
Comparison Context: Start with the comparative summary for this ecosystem at Best Open-Source Burmese & Myanmar LLMs.
Explore: Start with About, International Experience, Patents, and Talks & Media.

My Background

About Me

LinkedIn Google Scholar HuggingFace GitHub Medium

I am Dr. Wai Yan Nyein Naing, a Myanmar AI Applied Scientist and Senior Research Scientist at Robert Bosch, currently based in the United States.

My work spans Burmese language AI, large language models, agentic AI systems, multimodal learning, and enterprise AI deployment in safety-critical environments across Singapore and the USA.

As creator of Burmese GPT and Burmese-Coder-4B, I work to ensure the Burmese language is not left behind in the LLM era by building open, mobile-first, and practical AI for Myanmar language users.

Approximately 80% of Myanmar’s population is Myanmar-speaking, making Burmese-language AI utility especially important. This family of models is built for practical deployment in environments where connectivity is uneven, with focus on local workflows and low-resource device use.

LLMs & Fine-tuning

Agentic AI

Burmese NLP

Computer Vision

Multimodal AI

Enterprise AI

QLoRA / LoRA

MLX / GGUF

Selected Pages

Explore the work

Start here for my profile, timeline, patents, talks, and model releases.

Profile

About

Background, roles, publications, and the main public profile.

ProfileBackgroundResearchAbout

Open About

Career timeline

International Experience

Chronological career timeline across Myanmar, Malaysia, Singapore, and the United States.

TimelineBoschCareerExperience

View the timeline

Patents

Patents & Innovation

Patent list with publication numbers, assignees, and source links.

PatentsBoschInnovationResearch

Open the patent ledger

Talks

Talks & Media

Conference appearances, awards, and public recordings.

TalksMediaCommunityVideos

View talks and media

Releases

Model family

Burmese GPT, Padauk, Burmese-Coder-4B, the comparison guide, and the benchmark page.

ModelsBenchmarksReleasesLLMs

Compare the model family

Open About

Field Deployment Initiative

Bringing AI to Offline Environments

In Myanmar, internet access is uneven and can be disrupted, while demand for Burmese-language AI remains high. With about 33.4 million internet users and 63.3 million mobile cellular connections reported in early 2025, many workflows still require mobile-first, resilient approaches.

That is why this work focuses on on-device and offline-capable deployment patterns for practical use cases. Burmese GPT, Padauk, and Burmese-Coder-4B are aligned to support local experimentation, education, and daily tasks where continuous cloud access is not guaranteed.

The practical goal is not just faster models. It is reducing barriers to AI access for Myanmar language users, especially in mobile and low-resource contexts.

Contact For Deployment Details

Demonstration of customized 4B model inference running efficiently on Edge hardware.

Model Selection

Compare Open-Source Burmese AI Models

If you are choosing a model by use case, start with this practical map: foundation model, practical assistant layer, and coding model for Myanmar language workflows.

Burmese GPT

Burmese-first foundation

Open-source Burmese AI foundation model for Myanmar language adaptation, research, and downstream fine-tuning.

Open-source Burmese foundation model View comparison by use case

Padauk

Burmese AI assistant

Practical, agentic Burmese assistant for mobile-first, everyday tasks and low-resource user contexts.

Best for practical assistant workflows Compare with Burmese-Coder-4B and Burmese GPT

Burmese-Coder-4B

Coding model

Burmese coding model for developers, with practical use in low-resource devices and local workflows.

Burmese AI for coding tasks Context in Burmese AI landscape

Open the Full Comparison Guide

Open Source & Research

Featured Projects

Latest Publication

Burmese-Coder-4B featured on HackerNoon as a Burmese-focused code-generation model.

Burmese-Coder-4B project page | Read the HackerNoon feature

Foundational LLM

Burmese GPT

Open-source Burmese GPT foundation model for Myanmar language tasks, positioned for downstream adaptation into assistants, summarization, and local retrieval workflows.

Open-sourceBurmese FoundationNLPDeployment

Open-source Burmese foundation model (Burmese GPT)

Practical Assistant Layer

Padauk

Practical Burmese-first assistant layer for everyday tasks, tool workflows, and mobile-first deployment under constrained connectivity.

AgenticBurmese AssistantLow-resourceLocal Deployment

Burmese AI practical assistant (Padauk)

Education AI

Burmese Tutor

Local-first Burmese learning app that turns pasted study text into step-by-step lessons, with original screenshot and walkthrough video artifacts.

EducationOn-deviceOfflineOpen Source

Burmese tutor project page

Code Generation LLM

Burmese-Coder-4B

Open-source Burmese coding model for Myanmar developers, useful for code-heavy prompts, technical explanation, and local experimentation.

Code LLM4B ParamsMyanmar CodingLocal Deployment

Burmese coding model (Burmese-Coder-4B)

Benchmark Framework

burmese-coding-eval

A specialized multi-track benchmark and evaluation suite for Burmese programming assistants. Measures code correctness, linguistic quality, and cultural appropriateness of AI-generated code.

BenchmarkEvaluationPythonOpen Source

View Evaluation Framework

Agentic AI

rads-agent

An agentic AI system that converts any Kaggle dataset URL into automatic EDA, semantic SQL models, and natural-language data exploration — powered by multi-agent orchestration with DuckDB.

Agentic AIDuckDBMulti-AgentPython

View Project Details

Medical AI

AI Lung Infection Detection

Deep learning system for AI-powered lung infection and disease detection from medical imaging. Built for Myanmar clinical contexts with a Jupyter Notebook pipeline for research and education.

Computer VisionMedical AIDeep LearningMyanmar

View Project Details

Academic Research

Selected Publications

Research advances machine learning across time series forecasting, medical computer vision, anomaly detection, and Burmese language AI.

Survey
State of the Art Machine Learning Techniques for Time Series Forecasting: A Survey Nyein Naing, Wai Yan; Htike, Zaw Zaw Advanced Science Letters, 2015 Highly cited survey of ML methodologies for temporal data analysis
Computer Vision
Advances in Automatic Tuberculosis Detection in Chest X-Ray Images Wai Yan Nyein Naing, Zaw Z. Htike Signal & Image Processing: An International Journal Computer Vision applied to Medical CAD Systems for TB detection
Deep Learning
Glass breaks detection system using deep auto-encoders with fuzzy rules induction algorithm Wai Yan Nyein Naing, Zaw Zaw Htike, Amir Akramin Shafie International Journal of Advanced and Applied Sciences, 2019 Anomaly detection using deep generative models
Sequential Models
Real time end-to-end glass break detection system using LSTM deep recurrent neural network Wai Yan Nyein Naing, et al. Institute of Advanced Science Extension (IASE), 2019 End-to-end learning approach avoiding hand-crafted audio features

View All Publications

Industrial Impact

Patents & Innovation

11 patents filed at Robert Bosch GmbH spanning computer vision, AI diagnostics, multi-agent AI systems, terahertz sensing, and enterprise-grade AI in safety-critical environments.

View All 11 Patents

Recognition

Awards & Certifications

Bosch Inventor Award

Robert Bosch GmbH — 2023

Recognized for outstanding contributions to AI patent innovation within the Bosch Industrial AI division.

Innovation Award

Intl. Technology & Industrial Show — 2018

Recognized for AI Radiology Assistant (Myanmar) Application using Computer-Aided Diagnosis.

AI From the Data Center to the Edge

Intel Corporation — Credential ID: jst6kw4n

Optimization using Intel® Architecture for edge and data center AI deployment. Verify credential →

Gallery

Featured Media

Highlights from international tech conferences, startup press coverage, AI robotics projects, and innovation awards.

View Photo Gallery & Talks

Writing & Research

Latest Articles

#burmese tutor#ai4burmese#education ai

A Personalized Burmese Tutor That Runs On Your Device

burmese_tutor is an open-source Burmese tutor built to make learning more accessible. It turns pasted study material into step-by-step lessons, runs locally without internet in the offline setup, and combines practical tutoring behavior with a transparent AI4Burmese runtime.

Read the technical deep dive

#padauk#burmese gpt#gemma 4

Padauk Is a Burmese-First Agentic LLM for Low-Resource Devices

Padauk is not just another Burmese text generator. It is a Burmese-first agentic LLM based on Gemma 4 and specialized with a custom xLAM-format dataset for complex Burmese intent understanding and tool use, distributed through a primary Ollama-ready release and original fine-tuned weights on Hugging Face.

Read article

View All Articles

Reach Out

Get In Touch

Email

For research collaboration, speaking invitations, or professional inquiries.

waiyan.nn18@gmail.com

Connect professionally and discuss AI research, collaboration, or opportunities.

Connect on LinkedIn

GitHub

Browse source code, contribute to open-source Myanmar AI tools, and file issues.

View GitHub Profile

HuggingFace

Access all open-source models, datasets, and scripts published under WYNN747.

View WYNN747 Profile