Contributions

Ph.D. in Computer Science (AI Agents) with 13+ years of experience building and operating large-scale AI applications in production, including AI chatbots, LLM-based agents, and multi-agent orchestration.

Founded an early-stage startup, then designed and implemented Rakuten’s first in-house AI assistant platform, formalized as corporate IP and recognized by the CEO. The platform outperformed IBM Watson Discovery and Salesforce Search and now serves millions of users across e-commerce, mobile, and fintech.

Led global AI teams of up to 30 engineers to deliver 30+ production AI chatbots, reducing operational costs by ~30%. Primary inventor of patents granted in Japan and Taiwan, deployed at scale in Japan, spanning AI agent context sharing, enterprise knowledge grounding, privacy-preserving LLMs, AI quality improvement, and fresh food delivery optimization.

Led and architected the Rakuten Mobile AI Assistant (10M+ users), pioneering a multi-agent architecture and patented privacy-preserving LLM, earning the 2025 Japan IT Award for Customer Experience Innovation.

Building trustworthy, agentic AI systems that reason, act, and safely execute transaction-level operations at scale.

A. Granted patents deployed in production (登録済・本番稼働中の特許)

1. Nguyen The Minh, et al.: マルチ LLM エージェント環境における属性値の共有 , Granted in Japan (JP 7808651, Jan 21, 2026)

Enables multi-LLM orchestration via shared context memory.

2. Nguyen The Minh, et al.: 特定領域の知識を LLM に提供して回答させる , Granted in Japan (JP 7734169, Aug 27, 2025) and Taiwan (TWI911978B, Jan 11, 2026)

Enables LLMs to generate correct answers from private enterprise data.

3. Nguyen The Minh, et al.: 個人情報をマスクした質問文を用いて回答文を生成, Granted in Japan (JP 7667229, Apr 14, 2025) and Taiwan (TWI896396B, Sep 1, 2025).

Protects user privacy while preserving context for LLM responses.

4. Nguyen The Minh, et al.: Self NLU Improver, Granted in Japan (JP 7402277, Dec 12, 2023)

Improves AI model quality by generating effective training and test data.

5. Nguyen The Minh: 注文受付装置、注文受付方法、プログラム、ならびに、非一時的なコンピュータ読取可能な情報記録媒体 , Granted in Japan (JP 5976266, Jul 29, 2016)

Makes it easier to order fresh food by automatically aligning delivery dates.

6. Tran Tuan Tu, Nguyen The Minh: マルチ LLM エージェント環境における属性値の収集と属性値に基づくプロンプト生成, Granted in Japan (JP 7739557, Sep 5, 2025)

Enables LLMs to generate accurate responses for multi-AI agents by generating prompts with sufficient information.

B. Refereed journals

1. Nguyen Minh The: 実世界行動マイニングによる時系列行動ネットワークの構築～震災時の迅速な避難情報の共有に向けて～．博士論文，人工知能学会誌 28 巻 1 号（2013 年 1 月）．

2. The-Minh Nguyen, Takahiro Kawamura, Yasuyuki Tahara,  and Akihiko Ohsuga: Self-Supervised Capturing of Users’ Activities from Weblogs. International Journal of Intelligent Information and Database Systems,Vol.6, No.1, pp.61-76, InderScience Publishers, 2012 (pdf)

3. Nguyen Minh The, Takahiro Kawamura, Hiroyuki Nakagawa, Yasuyuki Tahara, Akihiko Ohsuga: Automatic Extraction and Evaluation of Human Activity using Conditional Random Fields and Self-Supervised Learning. Journal of the Japanese Society for Artificial Intelligence (JSAI), volume 26, no.1, pp.166-178, Special Issue on Agent, 2011 (in Japanese)

4. Toshiyuki Tanaka, The-Minh Nguyen, Hiroyuki Nakagawa,  Yasuyuki Tahara, and Akihiko Ohsuga: Opinion Extraction using Supervised Learning for Reputation Analysis System.  IEICE TRANSACTIONS on Information and Systems, Vol.J94-D, No.11, pp.1751-1761, 2011 (in Japanese)

5. Nguyen Minh The, Takahiro Kawamura, and Akihiko Ohsuga: Proposal Human Activity Estimation not Tweeted at Twitter : A Case Study of Returning Home in the Earthquake.  IEICE TRANSACTIONS on Information and Systems, J96-D(12), 2970-2978, 2013  (in Japanese)

C. Refereed international conference (view at DBLP)

1. The-Minh Nguyen, Takahiro Kawamura, Yasuyuki Tahara, and Akihiko Ohsuga: Toward Information Sharing of Natural Disaster: Proposal of Timeline Action Network. ICAART 2012, CCIS 358, pp. 145–157. Springer-Verlag Berlin Heidelberg 2013.

2. The-Minh Nguyen, Takahiro Kawamura, Yasuyuki Tahara, and Akihiko Ohsuga: Building a Time Series Action Network for Earthquake Disaster. In Proc. ICAART2012, volume 1, pp.100-108, Vilamoura, Algarve, Portugal, 6-8 February, 2012. In Cooperation with:  Portuguese Association for  Artificial Intelligence (APPIA), Spanish  Association of Artificial Intelligence (AEPIA), Association for the Advancement of Artificial Intelligence (AAAI). 

3. The-Minh Nguyen, Kenji Koshikawa, Takahiro Kawamura, Yasuyuki Tahara, and Akihiko Ohsuga: Building Earthquake Semantic Network by Mining Human Activity from Twitter. 2011 IEEE International Conference on Granular Computing, pp.496-501. Kaohsiung, Taiwan, November 8-10, 2011.

4. Isabel Shizu Miyamae Iwanaga, The-Minh Nguyen, Takahiro Kawamura, Hiroyuki Nakagawa, Yasuyuki Tahara, and Akihiko Ohsuga: Building an Earthquake Evacuation Ontology from Twitter. 2011 IEEE International Conference on Granular Computing, pp.306-311. Kaohsiung, Taiwan, November 8-10, 2011.

5. The-Minh Nguyen, Takahiro Kawamura, Yasuyuki Tahara, Akihiko Ohsuga: Capturing Users' Buying Activity at Akihabara Electric Town from Twitter. In Proc. ICCCI2010, part II, LNAI 6422, pp.163-171. Springer, Heidelberg, Kaohsiung, Taiwan, November 10-12, 2010.

6. Nguyen Minh The, Takahiro Kawamura, Hiroyuki Nakagawa, Yasuyuki Tahara, Akihiko Ohsuga: Automatic Mining of Human Activity and Its Relationships from CGM. In Proc. ICSOFT2010, volume 1, pp.285-292, Athens, Greece, 22-24 July, 2010.

7. Nguyen Minh The, Takahiro Kawamura, Hiroyuki Nakagawa, Yasuyuki Tahara, Akihiko Ohsuga: Automatic Mining of Human Activity Attributes from Weblogs. In Proc. IEEE/ACIS ICIS2010, Semantic, Social, and Collaborative Intelligence Session, pp.633-644, Yamagata, Japan, 18-20 August, 2010.

8. Nguyen Minh The, Takahiro Kawamura, Hiroyuki Nakagawa, Ken Nakayama, Yasuyuki Tahara, Akihiko Ohsuga: Human Activity Mining using Conditional Random Fields and Self-Supervised Learning. In Proc. ACIIDS2010,    part I, LNAI 5990, pp.140-149. Springer, Heidelberg. Hue City, Vietnam, 24 - 26 March, 2010.

9. Takahiro Kawamura, Nguyen Minh The, Akihiko Ohsuga: Building of Human Activity Correlation Map from Weblogs. In Proc. ICSOFT2009, volume 2, pp. 346-352. INSTICC Press, ISBN 978-989-674-010-8. Sofia, Bulgaria, 26-29 July, 2009.

D. Refereed international workshop

1. The-Minh Nguyen, Takahiro Kawamura, Yasuyuki Tahara, and Akihiko Ohsuga: Building a Timeline Network for Evacuation in Earthquake Disaster. The AAAI 2012 Workshop on Semantic Cities, pp.15-20, AAAI Publications. Toronto, Ontario, Canada, July 22-26, 2012 (pdf)

2. Nguyen Minh The, Takahiro Kawamura, Hiroyuki Nakagawa, Yasuyuki Tahara, Akihiko Ohsuga: Self-Supervised Mining of Human Activity from CGM. The 11th International Workshop on Knowledge Management and Acquisition for Smart Systems and Services (PKAW 2010), LNAI 6232, pp.61-72, Springer, Heidelberg. Daegu, Korea, August 30 – September 3, 2010.

E. Conference in Japan

1. グェン ミン テイ，越川兼地，川村隆浩，田原康之，大須賀昭彦： 震災時の効率的な避難のための行動推薦エージェントReady for 87%の提案（2）～時系列避難行動ネットワークの構築～．合同エージェントワークショップ＆シンポジウム2011．共催：日本ソフトウェア科学会，電子情報通信学会，情報処理学会，人工知能学会．後援：IEEE Computer Society Japan Chapter, 2009．

2. ミヤマエイワナガイサベルシズ，グェン ミン テイ，川村隆浩，中川博之，田原康之，大須賀昭彦：震災時の効率的な避難のための行動推薦エージェントReady for87%の提案（1）～避難計画オントロジーの設計～．合同エージェントワークショップ＆シンポジウム2011．共催：日本ソフトウェア科学会，電子情報通信学会，情報処理学会，人工知能学会．後援：IEEE Computer Society Japan Chapter, 2011．

3. 伊藤雅博，グェン ミン テイ，川村隆浩，中川博之，田原康之，大須賀昭彦：震災時の効率的な避難のための行動推薦エージェントReady for 87%の提案（3）～行動ネットワークを用いた避難行動の推薦～．合同エージェントワークショップ＆シンポジウム2011．共催：日本ソフトウェア科学会，電子情報通信学会，情報処理学会，人工知能学会．後援：IEEE Computer Society Japan Chapter, 2011．

4. グェン ミン テイ，伊藤雅博，川村隆浩，田原康之，大須賀昭彦：: Web からの実世界行動抽出による意味ネットワーク構築手法の提案. IEICE, Artificial Intelligence and Knowledge-Base Processing, vol. 111, no. 70, AI2011-8, pp. 39-44, Tokyo, 26th May, 2011.

5. 田中俊行， グェン ミン テイ， 中川博之， 田原康之，大須賀昭彦：シーソー型インタフェースを用いた評判比較システムの開発．情報処理学会，インタラクション 2011．

6. 田中俊行， グェン ミン テイ，中川博之， 田原康之，大須賀昭彦: 教師あり学習を用いたレビューサイトからの意見抽出. 合同エージェントワークショップ＆シンポジウム2010．共催：日本ソフトウェア科学会，電子情報通信学会，情報処理学会，人工知能学会．後援：IEEE Computer Society Japan Chapter, 2010．（学生奨励賞）

7. グェン ミン テイ，川 村隆浩，田原康之，大須賀昭彦： Twitterからの人間行動属性の自動抽出．IEICE・ 人工知能と知識処理研究会（Artificial Intelligence and Knowledge-Base Processing），一般講演（理論），pp.19-23 ．関 学東京キャンパス，2010．

8. グェン ミン テイ，川村隆浩，中川博之，田原康之，大須賀昭彦： CGMからの自己教師あり学習と条件付確率場を用いた人間行動マイニング．The 24th Annual Conference of the Japanese Society for Artificial Intelligence (JSAI2010). 社団法人 人工知能学会，2010.

9. グェン ミン テイ，川村隆浩，中川博之，田原康之，大須賀昭彦： Webからの自己教師あり学習を用いた人間行動マイニング．IEICE・人工知能と知識処理研究会（Artificial Intelligence and Knowledge-Based Processing）， Vol.109, No.386, pp.19-24. 「グリーンAI」および一般．関学東京キャンパス，2010．

10. グェン ミン テイ，川村隆浩，中川博之，中山健，田原康之，大須賀昭彦： 条件付確率場と自己教師あり学習を用いた行動属性の自動抽出と評価．合同エージェントワークショップ＆シンポジウム2009，pp. 468-475．共催：日本ソフトウェア科学会，電子情報通信学会，情報処理学会，人工知能学会．後援：IEEE Computer Society Japan Chapter, 2009．