Tomohiko Yokoyama

Currently an Applied Engineer at Sakana AI, where I build LLM agent systems with enterprise clients in the Japanese financial sector, taking projects from the initial proposal through PoC to enterprise solutions. I have delivered AI systems to Japanese enterprises and public-sector organizations, working with clients in Japanese and within engineering teams in Japanese / English. My background spans algorithm design, machine learning, and full-stack implementation. I hold a PhD in Information Science and Technology from the University of Tokyo.

ORCID Google Scholar Speaker Deck

Engineering Experience

Applied Research Engineer

(09/2025 - Present)
Sakana AI

R&D Engineer

(09/2022 - 09/2025)
Liquitous

ML Engineer, Data Scientist

(02/2022 - 02/2025)
Mirai Share

ML Engineer Intern

(08/2018 - 12/2018)
LegalOn Technologies

Research Experience

Assistant Technical Staff

(04/2022 - 09/2025)
The University of Tokyo

Research Visit

(01/2024 - 02/2024)
National University of Singapore

Research Scientist

(02/2023 - 08/2024)
CyberAgent

R&D Internship

(09/2020 - 09/2020)
NTT

Projects

Kajibuntan (Household Chore Division Application) GitHub Live Demo Article

We developed a web application using fair division algorithms for household chore allocation. Developed as part of a Code for Japan project. Featured on NHK (Japan's national public broadcasting).

PyMallows GitHub

This is a Mallows model sampling library written in Rust with Python bindings. Generating votes and rankings.

Education

PhD in Information Science and Technology, Mathematical Informatics

(10/2022 - 09/2025)
The University of Tokyo

Master of Information Science and Technology, Mathematical Informatics

(04/2020 - 03/2022)
The University of Tokyo

Bachelor of Engineering, Mathematical Engineering and Information Physics

(04/2016 - 03/2020)
The University of Tokyo

Publications

Journal

Reaction yield oscillates over reaction time in first-order chemical reactions Paper

Y. Harabuchi, T. Yokoyama, K. Katayama, S. Maeda, T. Oki, and S. Iwata

RSC Advances, 2026

Balancing fairness and high match rates in reciprocal recommender systems: A Nash social welfare approach Paper

Y. Tomita and T. Yokoyama

ACM Transactions on Recommender Systems, 2026

Asymptotic analysis of weighted fair division Paper

P. Manurangsi, W. Suksompong, and T. Yokoyama

Theoretical Computer Science, 2025

Virtual ligand-assisted optimization: A rational strategy for ligand engineering Paper

W. Matsuoka, T. Oki, R. Yamada, T. Yokoyama, S. Suda, C. M. Saunders, B. B. Skjelstad, Y. Harabuchi, N. Fey, S. Iwata, and S. Maeda

ACS Catalysis, 2024

Differentiating the yield of chemical reactions using parameters in first-order kinetic equations to identify elementary steps that control the reactivity from complicated reaction path networks Paper

Y. Harabuchi, T. Yokoyama, W. Matsuoka, T. Oki, S. Iwata, and S. Maeda

The Journal of Physical Chemistry A, 2024

Conference

Mechanism-informed learning for fair division Paper

A. Igarashi, R. Maruo, K. Takeuchi, and T. Yokoyama

In Proceedings of AAMAS 2026

Position fair mechanisms allocating indivisible goods Paper

R. Mahara, R. Mizutani, T. Oki, and T. Yokoyama

In Proceedings of AAAI 2026

* Oral presentation (top 5%)
Asymptotic analysis of weighted fair division Paper

P. Manurangsi, W. Suksompong, and T. Yokoyama

In Proceedings of IJCAI 2025

Probabilistic analysis of stable matching in large markets with siblings Paper GitHub

Z. Sun, T. Yokoyama, and M. Yokoo

In Proceedings of IJCAI 2025

Asymptotic existence of class envy-free matchings Paper

T. Yokoyama and A. Igarashi

In Proceedings of AAMAS 2025

Fair reciprocal recommendation in matching markets Paper GitHub

Y. Tomita and T. Yokoyama

In Proceedings of RecSys 2024

* Shortlisted for 6 best full paper candidates in RecSys'24
Kajibuntan: A house chore division app Paper GitHub

A. Igarashi and T. Yokoyama

In Proceedings of AAAI 2023

Other Involvements

Reviewing

Program Committee: AAAI 2027 / AAAI 2026