Overview

As Lead AI Scientist at SCB 10X, I collaborate with the Typhoon Team and SCBx R&D to democratize AI systems for low-resource languages and position Thailand as a leading AI innovation hub.

Key Achievements

• Founding member of the Typhoon Team, building it from the ground up into Thailand’s top AI research lab
• Led and launched multiple LLMs optimized for Thai, achieving over 100K downloads on Hugging Face and more than 30M API requests
• Actively contributed as both an individual contributor and supervisor across the majority of Typhoon Lab’s projects

Major Projects

• Typhoon2: text, vision, and audio multi-modality model (one of the first in SEA)
• Typhoon Audio2: one of the first end-to-end speech LLMs in SEA
• Typhoon T1: Southeast Asia’s first reasoning model
• Typhoon R1: the most advanced reasoning LLM tailored for Thai

Recognition

Won Techsauce innovation award 2024

Overview

As Senior AI & Software Engineer at SCB 10X, I worked with the Venture Building - Platform Team, crafting strategy and execution toward building the next Thai unicorn startup in the era of AI.

Key Achievements

• Main author of Typhoon, the most competitive Thai Large Language Model
• Designed a Thai knowledge evaluation system from scratch for LLMs
• Designed and implemented a continuous pretraining pipeline to adapt English-focused LLMs to Thai
• Published research: https://arxiv.org/abs/2312.13951
• Contributed to direction/theme and framework for AI project guidelines for the overall Venture Builder team
• Collaborated across all Venture Builder AI projects and initiatives

Impact

The work done during this period laid the foundation for what would become Typhoon Team, later established as Thailand’s premier AI research lab focused on language model development for Thai and regional languages.

Overview

As a Generative AI Engineer at a stealth LLM product, I worked remotely with the founding team to build a global LLM-powered B2C product serving over 1 million users.

Key Achievements

• Built a high-scale LLM infrastructure handling 2,000+ concurrent users and ~20 LLM calls per second
• Designed and developed LLM pipelines for real-time chat features
• Created innovative workarounds for open-source LLM limitations
• Deployed open-source LLMs at production scale while reducing costs by 10x
• Built and maintained infrastructure for high-throughput LLM inference

Technical Expertise

Specialized in solving scaling challenges for real-time LLM applications, optimizing inference speed and resource utilization, and implementing efficient caching and batching techniques for improved performance.

Overview

As an AI Software Engineer at SCB 10X, I worked part-time as Tech Lead & AI Engineer at spacely.ai, an AI interior design platform being developed with the goal of spinning off as a start-up under the venture builder team.

Key Achievements

• Led the technical development of the image generation model at spacely.ai
• Designed and implemented production-ready pipelines for AI image generation
• Established architectural foundations for scaling the platform
• Created optimization techniques for faster inference and better user experience
• Aligned technical implementation with business requirements and go-to-market strategy

Impact

Successfully laid the technical foundation for the image generation model and pipeline at spacely.ai, bringing it to production-level quality and preparing it for scale as an independent venture.

Overview

As an LLMs Consultant at Deeple, I was involved in prototyping Thai language LLMs and exploring integration strategies for leveraging large language models in business applications.

Key Contributions

• Led prototyping efforts for Thai language model adaptation
• Assessed business opportunities for LLM in the Thai market
• Explored fine-tuning strategies for specialized use cases

Impact

Helped the organization develop a roadmap for LLM era opportunity, identify viable business opportunities, and understand the technical requirements for successful implementation.

Overview

As Founder & CTO at Kunana Company Limited, I led all aspects of development across multiple products leveraging software and AI expertise to create innovative solutions for various markets.

Company Products

Vmagine (AI-Art Generator App)

• Mobile app using Image Generative AI for image/art creation
• Features: Image Generative AI, Full feature pro-grade image-editor
• Created company’s own model and an automatic personalized fine-tune system for AI avatars

KUU Creator

• Mobile app for online sellers to create ad/content images
• Features: In-house AI background removal technology, Smart AI Search system
• Reference: https://www.bangkokbanksme.com/en/kunana-kuu-creator-application

QQ

• Questionnaire platform for analyzing open-ended responses at scale
• Technology: Text Analytics / NLP AI
• Partner: Homerun consulting

AI Eyes

• AI solution predicting human perception of packaging
• Technology: Human Attention Prediction AI
• Partner: Homerun consulting

Client Projects

• Save my voice (2022): App helping hearing-impaired to train and see their voice.
  • Awards: Cannes Lions 2023, Adman 2023 and MAD STARS 2023.
• Missing Person Reporter (2021): Campaign for finding missing persons using face swap
  • Awards: Spikes Asia 2022, Adfest 2022, and Adman 2021.
• Visible-Track (2021): App translating visuals to sound for visually impaired
  • Awards: Adman Award 2021, Adfest 2022, and MAD STARS 2021.

Overview

As a Frontend Developer at Allianz Technology, I developed applications for insurance broker agents, focusing on performance optimization and process automation.

Key Achievements

• Improved the speed of a critical application feature by 10x
• Led a special project to create an automated solution for verifying insurance formulas within the app system
• Developed intuitive interfaces for complex insurance data
• Implemented performance optimizations for large data processing
• Collaborated with cross-functional teams to align technical solutions with business needs

Skills Applied

Frontend development, performance optimization, process automation, insurance domain knowledge, and cross-functional collaboration.

Overview

As a Founding Software Engineer at Deeple, I worked part-time to develop the MVP and first production version of the Deeple chatbot, establishing the technical foundation for the company’s conversational AI platform.

Key Achievements

• Led the development of an in-house NLP-AI model that being used in production for multiple year before the LLM era
• Built the initial MVP and first production version of the Deeple chatbot
• Designed natural language understanding components for Thai language
• Implemented core conversational flows and integration with back-end systems
• Created a scalable architecture to support future growth

Impact

The NLP models and architecture developed during this period proved to be robust and effective, continuing to power production systems years after initial development.