Built Thailand's leading AI research lab from the ground up. Led Typhoon models to achieve 1M+ downloads and 30M+ API requests, with enterprise deployment at Toyota Leasing, TDRI, PEA, VISAI, and Siriraj Hospital. Won Techsauce Innovation Award 2024 and AWS GAIA 2025 selection ($1M credits).

Built Thailand's leading AI research lab from scratch. Models outperform GPT-4o & Gemini on Thai-specific benchmarks (OCR, ASR, translation, instruction-following) with up to 156× cost reduction. Deployed at Toyota, hospitals, and government agencies.

Engineered high-scale LLM infrastructure for 1M+ users (2k concurrent, 20 calls/sec), developed real-time LLM pipelines, implemented custom workarounds on open-source LLMs during Llama 2 era, and deployed open-source LLMs at production scale.

Built production LLM infrastructure for 1M+ user B2C product. Achieved 2K+ concurrent users, 20 calls/sec with 10× cost reduction through optimization and custom workarounds.

Worked as early Tech Lead and AI Engineer at Spacely.ai, laying the foundation for the image generation model, pipeline, and achieve product-market fit, contributing to its development into fully spin-off as independent start-up with seed round funding.

Early Tech Lead establishing AI foundation (ControlNet, custom models). Platform scaled to 1M+ spaces, 120K+ users, spun off as independent startup.

Built an AI art generator app with custom diffusion models, pro-grade image editor, and personalized AI avatar system using automated fine-tuning.

Built AI art generator with custom Stable Diffusion models and automated avatar fine-tuning (K8s-based). ~4.6 App Store rating.

Developed an app using multiple speech AI technologies to verify various aspects of users' speech and help them practice.

First speech training platform for hearing-impaired using AI. Won Cannes Lions 2023, Adman 2023, MAD STARS 2023.

Developed an AI solution using computer vision and attention models to predict human perception of packaging, featuring heat maps and design recommendations.

Computer vision solution for packaging design optimization. Deployed with multiple FMCG corporate clients in Thailand.

Developed a platform using NLP (topic modeling, clustering) for automated analysis of open-ended questionnaire responses.

NLP platform for automated questionnaire analysis at scale. Topic modeling and semantic clustering for qualitative insights.

Leveraged face swap technology for a social impact campaign by transferring single-image face swaps of missing person onto key opinion leaders (KOLs) to make the campaign more memorable.

Real-time face swap AI for social impact. Won Spikes Asia 2022 Silver, Adfest 2022 Silver, Adman 2021. Increased missing person awareness and reports.

Developed an app using real-time AI visual detection to translate visuals to sound.

Computer vision accessibility tool (ResNet, CenterNet, TransNet). Won Adman 2021, Adfest 2022 Bronze, MAD STARS 2021 Bronze.

Built a mobile app with in-house AI background removal and Smart AI Search system, developing the complete pipeline from data labeling to model training and deployment.

Built in-house AI background removal competitive with PhotoRoom/Remove.BG (2020). ~4.7 App Store rating. Complete ML pipeline from labeling to production.