Thodsaporn Chay-intr
Chief AI Officer at iApp Technology, Bangkok
I build Thai-language AI at iApp, where I lead the AI and R&D departments and still write code most days. My team's models include OpenThai Legal, OpenThaiGPT, ChindaLLM, and ChindaMT; together with our speech and vision systems, they have served more than 70 million API calls. I work across the whole stack, from post-training with SFT, DPO, and GRPO down to vLLM and TensorRT-LLM serving on our own GPUs.
I hold a doctorate from Institute of Science Tokyo (formerly Tokyo Institute of Technology), where I worked on character-based word segmentation in the Okumura–Funakoshi Lab. These days I also advise the Big Data Institute on inference serving for ThaiLLM, Thailand's sovereign language model programme.
Some of this is easy to check: Tech Collective named iApp a top-five AI startup in Thailand in 2026, our face anti-spoofing holds an iBeta Level 1 certification, and the models and papers are on open.iapp.co.th and arXiv.
Work I'd point you to first
OpenThai 2.0 Legal (2026) — an open-weight Thai legal language model, built with NVIDIA, BDI, and AIEAT on Nemotron-3-Nano-30B. Legal Thai is formal, precise, and badly served by general models; this one is free to download and run.
OpenThaiGPT 1.6 & R1 (2025) — Thai-centric reasoning models. R1 is built on DeepSeek-R1-Distill-32B and beats DeepSeek-R1-70B on Thai reasoning benchmarks at half the size: 56.7 against 33.3 on AIME24-TH. There is a paper, and the weights are public: 1.6-72B, R1-32B.
CIBbot (2026) — a legal-advice and incident-reporting assistant we built for the Royal Thai Police's Central Investigation Bureau. It runs fully air-gapped over seven bureau archives and answers with 96.6% retrieval accuracy on a 208-item legal evaluation. It was launched at CIB headquarters with the National Research Council, and you can try it.
one.iapp.co.th and open.iapp.co.th (2026) — iApp's two public faces. One sells our text, voice, and vision APIs; open gives models away under Apache-2.0, including Chinda LLM 4B, ChindaMT, and the ChindaTTS voices.
Certified biometric liveness (2024) — dual-stage face anti-spoofing that passed iBeta Level 1 Presentation Attack Detection testing with a 0% attack presentation classification error rate. We have since extended it to passive deepfake screening for banking.
ThaiLLM (2025–) — I advise the Big Data Institute on serving infrastructure for Thailand's national 8B and 30B models, drawing on our own vLLM and TensorRT-LLM benchmarks on H100.
News
- Jul 2026 — Tech Collective named iApp a top-five AI startup in Thailand.
- Jul 2026 — We released OpenThai 2.0 Legal open-weight, built with NVIDIA, BDI, and AIEAT on Nemotron-3-Nano-30B.
- Jun 2026 — CIBbot launched at CIB headquarters with the Royal Thai Police and the National Research Council.
- May 2026 — ChindaMT released: instruction-following Thai–English translation, open-weight.
- Dec 2025 — SiamGPT preprint: quality-first fine-tuning for stable Thai text generation.
- Sep 2025 — ChindaGo shipped: any open-weight LLM running fully on-device on iOS.
- Jun 2025 — Promoted to Chief AI Officer at iApp.
- May 2025 — Co-authored Thailand's national OCR performance-testing standard, published by ETDA and drafted with NECTEC.
- Apr 2025 — The OpenThaiGPT 1.6 & R1 paper went up on arXiv.
- Mar 2025 — Chinda LLM 4B released under Apache-2.0 on open.iapp.co.th.
Selected publications
- SiamGPT: Quality-First Fine-Tuning for Stable Thai Text Generation. arXiv:2512.19455, 2025.
- OpenThaiGPT 1.6 and R1: Thai-Centric Open Source and Reasoning Large Language Models. arXiv:2504.01789, 2025.
- LLaVAC: Fine-tuning LLaVA as a Multimodal Sentiment Classifier. arXiv:2502.02938, 2025.
- Extreme Fine-tuning: A Novel and Fast Fine-tuning Approach for Text Classification. EACL, 2024. Paper, code.
- LATTE: Lattice ATTentive Encoding for Character-based Word Segmentation. Journal of Natural Language Processing, 2023. Paper, code.
The full list, with code links, is on the publications page.