# Parinthapat Pengpun (Parin)

> Parinthapat (Parin) Pengpun: AI undergrad at Carnegie Mellon and undergraduate researcher at CMU NeuLab, working on teaching LLMs and agents complex tasks with less compute and supervision. Creator of Typhoon Vision, Thailand’s first vision-language model.

Parinthapat Pengpun, who goes by Parin (handle: parinzee), is an undergraduate studying Artificial Intelligence at Carnegie Mellon University's School of Computer Science, and an undergraduate researcher at CMU's NeuLab (Language Technologies Institute), advised by Prof. Graham Neubig, since Aug 2025. He is an SCBX Young Tech Scholar (full scholarship). He is from southern Thailand, went to high school at Bangkok Christian International School in Bangkok, is now based in Pittsburgh, PA, and speaks Thai and English. Seeking Summer 2027 research internships.

## Quick answers

**Who is Parinthapat Pengpun?**
Parinthapat (Parin) Pengpun: AI undergrad at Carnegie Mellon and undergraduate researcher at CMU NeuLab, working on teaching LLMs and agents complex tasks with less compute and supervision. Creator of Typhoon Vision, Thailand’s first vision-language model.

**What does Parin research?**
Making it cheaper to teach LLMs and agents complex tasks: synthetic data and environments, curricula, and reward design, especially when compute, data, or supervision is scarce.

**Where does Parin study and do research?**
B.S. Artificial Intelligence, Carnegie Mellon University (School of Computer Science). Undergraduate Researcher at CMU LTI · NeuLab, advised by Prof. Graham Neubig, since Aug 2025.

**What is Typhoon Vision?**
Thailand's first vision-language model (8B parameters). Parin led and built it at SCB10X (Mar 2024 – Jan 2025), with support from the Typhoon team, while in 11th grade. At release it beat Qwen2-VL and Llama-3.2 on Thai visual question answering.

**What are Parin's main papers?**
Think Before You Link: Rarity, Reasoning, and Retrieval in Multilingual Entity Linking (EMNLP 2026, first author); On Creating an English-Thai Code-switched Machine Translation in Medical Domain (Findings of EMNLP 2024, co-first author); Seed-Free Synthetic Data Generation Framework for Instruction-Tuning LLMs: A Case Study in Thai (ACL SRW 2024, first author); SICAR at RRG2024: GPU Poor's Guide to Radiology Report Generation (BioNLP @ ACL 2024, co-first author).

**What industry experience does Parin have?**
Multimodal Junior Scientist at SCB10X (Mar 2024 – Jan 2025); Data Science Intern at Siam Commercial Bank (Jun – Jul 2026), doing uplift modeling (causal ML); AI/ML Engineer at Cognitus (acq. by IBM) (2024 – 2025), building self-verifying LLM agents.

**What honors has Parin received?**
Heidelberg Laureate Forum (2026): Thailand’s sole computer-science representative, nominated by NSTDA; one of 200 young researchers worldwide; HackIllinois (2026): three awards for OpenReality: First Place, Best AI Inference (Modal track); MLH Best Use of Google Gemini API; Most Creative; SCBX Young Tech Scholar (2025): full scholarship; sole recipient from Thailand in the 2025 cohort; AI Thailand Hackathon (2024): 1st Runner-Up with a multimodal Thai tourism agent, later the MAHASAMUT paper; SCB10X Bangkok AI Hack (2023): 1st place (Track 1), plus the VISTEC and AWS bounties; Super AI Engineer, Season 3 (2023): Gold Medal, the program’s top distinction; one of seven Gold Medalists, and the first high schooler to win it; AI Builders 2 (2022): 1st place and Best NLP project.

**Is Parin looking for internships?**
Seeking Summer 2027 research internships. CV: https://parinzee.github.io/cv.pdf.

**How can I contact Parin?**
Use the email button on https://parinzee.github.io/ (the address is not published in plain text), or LinkedIn https://www.linkedin.com/in/parinthapat-pengpun/, or X https://x.com/parinzee.

## Pages

- [Home](https://parinzee.github.io/): profile and an interactive terminal session (whoami, /papers, /experience, /honors, /hire)
- [About](https://parinzee.github.io/about): a Hugging Face-style "model card" with bio, work experience, every paper, awards, and acknowledgements
- [Research](https://parinzee.github.io/research): every project, with links to the papers
- [Blog](https://parinzee.github.io/blog) · [RSS](https://parinzee.github.io/rss.xml) · [CV (PDF)](https://parinzee.github.io/cv.pdf)

## Experience

- **Undergraduate Researcher, CMU NeuLab** (Aug 2025 – now): Research on teaching LLMs and agents complex tasks with less compute and supervision, advised by Prof. Graham Neubig. First-author paper at EMNLP 2026 on multilingual entity linking.
- **Teaching Assistant, 15-213, CMU** (Fall 2026): Introduction to Computer Systems: office hours and recitations.
- **Data Science Intern, Siam Commercial Bank** (Jun – Jul 2026): Built a causal uplift model that ranks customers by how much a campaign changes what they do, not just how likely they are to buy. In offline evaluation (Jan–Feb 2026 data) it was estimated to deliver 16% more conversions at the same send budget than the bank’s existing ranking.
- **AI/ML Engineer, Cognitus (acq. by IBM)** (2024 – 2025): Built self-verifying LLM agents (planner, executor, verifier) that author workflows across 100+ custom node types on the company’s workflow platform.
- **Multimodal Junior Scientist, SCB10X** (Mar 2024 – Jan 2025): Led and built Typhoon Vision (8B), Thailand’s first vision-language model: data pipeline, training, and release, with support from the Typhoon team. At release it beat Qwen2-VL and Llama-3.2 on Thai visual question answering. Co-author on Typhoon 2.
- **Synthetic Data Research Intern, VISTEC** (May – Aug 2024): Generated 100K+ Thai instruction-tuning examples grounded in Wikipedia; a curated 5K subset beat training on all 100K. Adopted by two national AI programs, WangchanX and AI Singapore.

## Papers

* = equal contribution

- Think Before You Link: Rarity, Reasoning, and Retrieval in Multilingual Entity Linking. P. Pengpun, S. Khanuja, G. Neubig. *EMNLP 2026*, 2026. first author. https://arxiv.org/abs/2609.10745
- When Better Teachers Don't Make Better Students: Revisiting Knowledge Distillation for CLIP Models in VQA. P. Tuchinda, P. Pengpun, R. Chumpu, P. Payoungkhamdee, S. Nutanong, P. Limkonchotiwat. *MRAC @ ACM MM 2026*, 2026. 2nd author. https://arxiv.org/abs/2511.17886
- On Creating an English-Thai Code-switched Machine Translation in Medical Domain. P. Pengpun*, K. Tiankanon*, A. Chinkamol, J. Kinchagawat, et al. *Findings of EMNLP 2024*, 2024. co-first author (equal contribution). https://aclanthology.org/2024.findings-emnlp.351/
- Typhoon 2: A Family of Open Text and Multimodal Thai Large Language Models. with K. Pipatanakul, P. Manakul, N. Nitarach, et al. *arXiv*, 2024. co-author. https://arxiv.org/abs/2412.13702
- Seed-Free Synthetic Data Generation Framework for Instruction-Tuning LLMs: A Case Study in Thai. P. Pengpun, C. Udomcharoenchaikit, W. Buaphet, P. Limkonchotiwat. *ACL SRW 2024*, 2024. first author. https://aclanthology.org/2024.acl-srw.50/
- SICAR at RRG2024: GPU Poor's Guide to Radiology Report Generation. K. Udomlapsakul*, P. Pengpun*, T. Saengja, K. Veerakanjana, et al. *BioNLP @ ACL 2024*, 2024. co-first author (equal contribution). https://aclanthology.org/2024.bionlp-1.55/
- MAHASAMUT: AI-Powered Thai Tourism using Multimodal Agents. with N. Ketui, K. Boonmeeprakob, P. Khlaisamniang, T. Supnithi. *LaCATODA @ PRICAI 2024*, 2024. 2nd author. https://ceur-ws.org/Vol-3862/paper5.pdf
- Can General-Purpose Large Language Models Generalize to English-Thai Machine Translation? with J. Chiaranaipanich, N. Hanmatheekun, et al. *arXiv*, 2024. co-author. https://arxiv.org/abs/2410.17145
- Elucidating the Relationship Between Climate Change and Poverty using Graph Neural Networks, Ensemble Models, and Remote Sensing Data. P. Pengpun, A. Salatiello. *CCAI @ NeurIPS 2023*, 2023. first author. https://www.climatechange.ai/papers/neurips2023/117
- Cross-Lingual Data Augmentation For Thai Question-Answering. P. Pengpun, C. Udomcharoenchaikit, W. Buaphet, P. Limkonchotiwat. *GenBench @ EMNLP 2023*, 2023. first author. https://aclanthology.org/2023.genbench-1.16/

## Honors

- Heidelberg Laureate Forum (2026): Thailand’s sole computer-science representative, nominated by NSTDA; one of 200 young researchers worldwide
- HackIllinois (2026): three awards for OpenReality: First Place, Best AI Inference (Modal track); MLH Best Use of Google Gemini API; Most Creative
- SCBX Young Tech Scholar (2025): full scholarship; sole recipient from Thailand in the 2025 cohort
- AI Thailand Hackathon (2024): 1st Runner-Up with a multimodal Thai tourism agent, later the MAHASAMUT paper
- SCB10X Bangkok AI Hack (2023): 1st place (Track 1), plus the VISTEC and AWS bounties
- Super AI Engineer, Season 3 (2023): Gold Medal, the program’s top distinction; one of seven Gold Medalists, and the first high schooler to win it
- AI Builders 2 (2022): 1st place and Best NLP project

## Other projects

- **Resnik** (CMU NASA SUITS team, 2026): Built the team’s end-to-end AI voice assistant, designed to support astronauts during spacewalk (EVA) simulations.
- **OpenReality** (HackIllinois 2026): A phone browser streams a room into a live 3D map. Parin built the autonomous agent that decides what to look for in the space while mapping continues. Three awards. https://devpost.com/software/real-eyes-mra8n6

## Mentors and acknowledgements

The people Parin credits for getting him here, with his one-line thanks in his own words. (P’ is a Thai honorific for an older peer.)

- [Graham Neubig](https://www.phontron.com/), Associate Professor, CMU Language Technologies Institute: "Advises my research at NeuLab."
- [Sarana Nutanong](https://vistec.ac.th/faculty-adjunct-professors-ist-sarana-nutanong), Associate Professor, VISTEC School of Information Science and Technology: "Brought me into VISTEC and introduced me to P’Ping."
- [Peerat Limkonchotiwat](https://mrpeerat.github.io/) (P’Ping), Research Fellow, AI Singapore · PhD, VISTEC: "The person who taught me how to do research."
- [Nongnuch Ketui](https://aiat.or.th/commitee/) (Prof. Nuch), Computer Science, Rajamangala University of Technology Lanna (Nan) · AIAT Vice President: "Helped me into the AI industry after Super AI Engineer."
- [Thepchai Supnithi](https://aiat.or.th/commitee/) (Prof. Thepchai), Senior Researcher, NECTEC · AIAT President: "Opened the door to the AI industry for me in high school."
- [Simran Khanuja](https://simran-khanuja.github.io/), PhD candidate, CMU Language Technologies Institute: "Brought me into NeuLab; my mentor on Think Before You Link."
- [Weerayut Buaphet](https://weerayutbu.github.io/) (P’Fame), VISTEC NRL Lab: "Helped me at AI Builders and introduced me to Prof. Sarana."
- [Can Udomcharoenchaikit](https://scholar.google.com/citations?user=dgga01IAAAAJ) (P’Can), NLP researcher, VISTEC: "Taught me research at VISTEC."

## Writing

- [A Reflection on 18 Years of Life](https://parinzee.github.io/blog/a-reflection-on-18-years-of-life) (2026-01-04): What truly makes a life enjoyable? How does one achieve their goals without losing themselves? What's my life purpose?

## Recent news

- Sep 2026: Represented Thailand at the Heidelberg Laureate Forum. Selected by NSTDA; one of 200 young researchers worldwide.
- Sep 2026: Paper accepted to EMNLP 2026. Think Before You Link, first author, main conference.
- Jan 2026: New post: A Reflection on 18 Years of Life. On chasing goals without losing yourself.
- Aug 2025: Joined CMU NeuLab. Undergraduate researcher, advised by Graham Neubig.
- Jul 2025: Named an SCBX Young Tech Scholar. Full scholarship; sole recipient from Thailand in 2025.

## Profiles

- Google Scholar: https://scholar.google.com/citations?user=1gbFjCwAAAAJ
- GitHub: https://github.com/parinzee
- Hugging Face (models): https://huggingface.co/parinzee
- ORCID: https://orcid.org/0009-0004-5562-327X
- LinkedIn: https://www.linkedin.com/in/parinthapat-pengpun/
- X: https://x.com/parinzee