Portrait of Ken Kato

Tokyo · Open to new opportunities

Hi, I'm Ken.

I'm a software engineer with a background in AI and NLP research based in Tokyo. I build tools at the intersection of language models, robotics, and political science — currently working on KOKKAI DOC, an LLM‑powered platform making Japanese parliamentary politics more transparent.

01 · Experience

Where I've worked and researched

Across academic labs, startups and Fortune-500 teams — in research, full-stack and ML engineering.

  1. Tokyo, Japan · Dec 2025 — Present

    SWE · VETA Inc.

    • Full stack
    • Infra
    • GCP
    • Terraform
    • Led the development of a new B2B SaaS platform allowing clients to configure conjoint surveys and gain insight from the collected data
    • The platform was developed with a Next.js frontend, fastapi backend containerized with docker. Using firestore DB. Was deployed on GCP using github workflows.
    veta.co.jp

  2. Tokyo, Japan · Aug 2025 — Dec 2025

    Data Scientist · Digital Agency @ Japanese Government

    • AI
    • LLM
    • NLP
    • Contributed to the development of “Gennai”, AI-driven SaaS application,, applying LLMs to government-specific tasks
    • Developed AI-driven micro-services in python, deployed on Google Cloud function that specialize in domains of government operations
    • Led the development of a monitoring system that monitors and alerts about the operations of the AI-driven micro-services. measuring both quantitative and qualitative metrics, such as endpoint errors and the quality of LLM responses. Mainly working with AWS CDK
    • Quit after being requested to stop running my website, kokkaidoc.com which I have been open about during the application process
    digital-agency-news.digital.go.jp

  3. Munich, Germany · May 2024 — Aug 2024

    AI/Robotics Researcher · Learning Systems Robotics Lab at TUM

    • AI
    • LLM
    • Robotics
    • ROS2
    • Mujoco
    • RL
    • Research
    • NLP
    • Worked towards integrating LLMs into robotic controls to enable semantically nuanced behaviour of human assisting robots
    • Developed a simulation environment for the Stretch 3 platform to test out the deployment of RT1 model on the platform
    • Integrated RT1 model developed by Google DeepMind on the physical Stretch 3 platform to successfully conduct picking up tasks
    • Designed soft-material robotic grippers and iterated on the design to successfully create a motor actuated soft gripper
    GitHubGitHub

  4. Toronto, Canada · Oct 2023 — Apr 2024

    NLP Researcher · University of Toronto

    • NLP
    • LLM
    • Japanese Politics
    • Political Science
    • Research
    • Political Methodology
    • Publication
    • PolMeth2024
    • Worked as a NLP researcher under co-supervision of Prof. Raeid Saqur and Dr. Christopher Cochrane to work on a paper exploring the use of LLM in the analysis of parliamentary speeches of Japanese legislators to predict their ideological stances in regards to various political topics.
    • Fine-tuned a BERT model to classify speech segments into opinion, factual, descriptive, other sentence types
    • Conducted statistical analysis on the retrieved embeddings to explain political stances of legislators
    • Published paper 'L(u)PIN: LLM-based Political Ideology Nowcasting' on arxiv.
    • Presenting a poster in the PolMeth 2024 conference at UC Riverside.
    arXivPoster Presentation at PolMeth 2024

  5. Toronto, Canada · Oct 2023 — Apr 2024

    NLP Researcher · Vector Institute

    • NLP
    • LLM
    • Finance
    • Prompt Engineering
    • Publication
    • Research
    • We introduce and make publicly available the NIFTY Financial News Headlines dataset, designed to facilitate and advance research in financial market forecasting using large language models (LLMs). This dataset comprises two distinct versions tailored for different modeling approaches: (i) NIFTY-LM, which targets supervised fine-tuning (SFT) of LLMs with an auto-regressive, causal language-modeling objective, and (ii) NIFTY-RL, formatted specifically for alignment methods (like reinforcement learning from human feedback (RLHF)) to align LLMs via rejection sampling and reward modeling. Each dataset version provides curated, high-quality data incorporating comprehensive metadata, market indices, and deduplicated financial news headlines systematically filtered and ranked to suit modern LLM frameworks. We also include experiments demonstrating some applications of the dataset in tasks like stock price movement and the role of LLM embeddings in information acquisition/richness. The NIFTY dataset along with utilities (like truncating prompt's context length systematically) are available on Hugging Face at this https URL.
    arXiv

  6. Toronto, Canada · May 2023 — Apr 2024

    Software Engineer Co-op Intern · Hellofresh Canada

    • React
    • Frontend Development
    • Web
    • Agile Workflow
    • Cypress
    • React Testing Library
    • Honeycomb
    • Google Analytics
    • Statsig
    • Optimizely
    • TypeScript
    • NextJS
    • Accelerated the development of the rapidly expanding web platform for HelloFresh and its sub-brands including ChefsPlate, Factor, GreenChef, and EveryPlate, enhancing global user access.
    • Worked with a NextJS and TypeScript stack to develop new features and improve existing ones, including the implementation of a new user onboarding flow and the integration of a new payment gateway.
    • Engineered and executed end-to-end tests with Cypress and developed unit tests with React Testing Library to ensure software performance.
    • Established monitoring protocols using Honeycomb, implemented Google Analytics tracking, and led A/B testing initiatives with Statsig and Optimizely to optimise user experience and operational efficiency.

  7. Toronto, Canada · May 2022 — Feb 2023

    Censorship Researcher · Citizen Lab at University of Toronto

    • Censorship
    • NLP
    • Web Scraping
    • Research
    • Publication
    • Working as a research fellow at the Citizen Lab in the University of Toronto.
    • Received funding worth more than $6000 as a ESROP(Engineering Science Research Opportunity Program) fellow from the Division of Engineering Science and Citizen Lab.
    • Contributed to the research done by Dr. Jeffrey Knockel, analyzing the censorship implementation in search engines operated by Chinese ISPs.
    • Published the work done on the Citizen Lab website.
    • Publication featured on the New York Times
    Citizen LabNew York Times

  8. Tokyo, Japan · May 2020 — Jun 2021

    Software Engineer Intern · Kozo Keikaku Engineering Inc.

    • SOLIDWORKS
    • PARTICLEWORKS
    • Web Dev
    • Frontend Development
    • Backend Development
    • NLP
    • Deep Reinforcement Learning
    • C++
    • Flask
    • HTML
    • CSS
    • JS
    • Mecab
    • Gensim
    • Tensorflow
    • Structural analysis using the SOLIDWORKS Simulation as well as fluid flow analysis using SOLIDWORKS Flow Simulation and Particle Works
    • Implementation of physical models using C++
    • Full-stack product development in flask and HTML, CSS and JS
    • Contributing to the NLP product development of the division, using Mecab and gensim
    • Deep reinforcement learning experience, implementing models such as A2C, REINFORCE and DQN in tensorflow.

  9. Tokyo, Japan · Aug 2020 — Aug 2020

    Business Consultant Intern · PwC

    • Consulting
    • Business
    • Business Strategy
    • Worked as a business strategy consultant intern at PwC Japan
    • Advised real-life clients on their future business strategy and operations
    • Worked together with a team of interns to put together a final pitch to the clients

  10. Dusseldorf, Germany · Jun 2016 — Feb 2019

    U14 Football Coach · SC West Düsseldorf e.V.

    • Leadership
    • Football
    • Coaching
    • Teamwork
    • Communication
    • Worked as a U14 football coach to train a team composed of Japanese and German players.

  11. Dusseldorf, Germany · Jun 2017 — Jul 2017

    Heavy Machinery Student Intern · SIEMENS AG

    • Welding
    • Heavy Industry
    • Metal Works
    • Mechanical Engineering
    • As a student intern at SIEMENS AG, I was exposed to first hand experience of mechanical engineering by assembling mechanical parts and welding; which I enjoyed the most. This was a precious experience for me because this was what convinced me to pursue engineering in university.

02 · Projects

Things I've built

Selected projects, side experiments, and the tools I'm currently shipping.

Jun 2023 — Present

KOKKAI DOC

Featured
  • NLP
  • LLM
  • AI
  • Data Science
  • Data Visualization
  • Finetuning
  • React
  • Political Methodology
  • Startup
  • Founder
  • Frontend Development
  • Backend Development
  • KOKKAI DOC is a platform that I founded which aims to make Japanese politics more transparent to the public through the use of AI/LLMs and data science
  • The platform uses a fine-tuned LLM to analyze the speeches of Japanese legislators and predict their ideological stances in regards to various political topics
  • The platform also provides a intuitive GUI enabling users to look up the speeches of representatives from various electoral districts and compare their stances on various topics
  • I have also scraped the voting data of parliamentary representatives and visualized them in a more intuitive way to help users understand the voting patterns of the representatives and parties
Visit kokkaidoc.com

Nov 2024 — Nov 2024

AI Website Builder

Featured
  • LLM
  • AI
  • React
  • Frontend Development
  • Backend Development
  • Node.JS
  • TypeScript
  • NextJS
  • TailwindCSS
  • MongoDB
  • Developed an AI-powered web application enabling non-technical users to create and edit HTML/CSS websites through natural language interactions with an AI assistant.
  • Frontend: Built with Next.js, TypeScript, and TailwindCSS for a responsive and modern UI.
  • Backend: Developed using Node.js with Express and TypeScript, acting as an intermediary between the frontend and OpenAI's API.
  • Database: Utilized MongoDB to store user sessions, including generated HTML/CSS, AI interactions, and metadata.
  • Users can create accounts, authenticate via JWT, and store session histories.
  • Ensured scalable deployment using Docker Compose, containerizing all components for seamless setup and execution.

Jun 2021 — Jul 2021

Diary CRUD Application with React and Django

  • React
  • Django
  • CRUD
  • Web Development
  • Python
  • Backend Development
  • Frontend Development
View Repository

Nov 2020 — Jan 2021

School Reunion CRUD Social Media Platform using Django, HTML, CSS, JS

  • HTML/CSS/JS
  • Django
  • CRUD
  • Web Development
  • Python
  • Backend Development
  • Frontend Development
View Repository

03 · Publications

Research & writing

Peer-reviewed and open-access research in NLP, political methodology, and internet freedom.

Sep 2024 — Apr 2025

KOKKAI DOC: An LLM-driven framework for scaling parliamentary representatives

  • JSQPS 2025
  • AI
  • LLM
  • NLP
  • Political Science
  • Publication
  • Political Methodology

This paper introduces an LLM-driven framework designed to accurately scale the political issue stances of parliamentary representatives. By leveraging advanced natural language processing techniques and large language models, the proposed methodology refines and enhances previous approaches by addressing key challenges such as noisy speech data, manual bias…

Read on arXiv

Nov 2023 — Apr 2024

NIFTY Financial News Headlines Dataset

  • NLP
  • LLM
  • Finance
  • Prompt Engineering
  • Publication

We introduce and make publicly available the NIFTY Financial News Headlines dataset, designed to facilitate and advance research in financial market forecasting using large language models (LLMs). This dataset comprises two distinct versions tailored for different modeling approaches: (i) NIFTY-LM, which targets supervised fine-tuning (SFT) of LLMs with an a…

Read on arXiv

Oct 2023 — Apr 2024

L(u)PIN: LLM-based Political Ideology Nowcasting

  • AI
  • LLM
  • NLP
  • Political Science
  • Publication
  • Political Methodology
  • PolMeth2024

The quantitative analysis of political ideological positions is a difficult task. In the past, various literature focused on parliamentary voting data of politicians, party manifestos and parliamentary speech to estimate political disagreement and polarization in various political systems. However previous methods of quantitative political analysis suffered…

Read on arXiv

May 2022 — Feb 2023

Missing Links A comparison of search censorship in China

  • Censorship
  • NLP
  • Data Science
  • Web Scraping
  • New York Times
  • Research
  • Publication

Across eight China-accessible search platforms analyzed — Baidu, Baidu Zhidao, Bilibili, Microsoft Bing, Douyin, Jingdong, Sogou, and Weibo — we discovered over 60,000 unique censorship rules used to partially or totally censor search results returned on these platforms.

Read paperNew York Times

04 · Credentials

Education & awards

Education

Sep 2019 — Apr 2025 · Toronto, Canada

University of Toronto - Engineering Science

I graduated from the University of Toronto with a Bachelor of Applied Science in Engineering Science with a specialization in Machine Intelligence.

Awards & Funding