Welcome to My Portfolio

Hello! I'm Danmeng CAI, an AI Engineer at FCNT (Lenovo Group) in Kanagawa, Japan. I work with the Motorola global team to create AI-powered features that enhance smartphone experiences

I hold an MS in Informatics (NLP specialization) and an MA in International Area Studies from the University of Tsukuba.

My background spans AI product management, LLM development, and international collaboration, allowing me to connect technical expertise with global perspectives to deliver innovative and user-centered AI solutions.

Projects

localizationof AI features for Motorola smartphones

Localizing Moto AI features in Japanese market

After joining Lenovo, I have been working on localizing AI features for Motorola smartphones to better suit the Japanese market. I collaborate with cross-functional teams across China, the US, India, and Brazil to adapt AI functionalities, ensuring they meet local user preferences and cultural nuances. I lead the testing team in Japan, providing weekly reports on localization progress and coordinating feedback to enhance feature performance.

See the news about Moto AI

FamLink - system that connects and supports families

As part of the FCNT team, I joined the Motorola Hackathon 2025 in Xiamen and organized the development of FamLink, a proof-of-concept AI service that connects elderly users’ devices, secure family databases, and conversational AI agents to enable remote wellbeing monitoring. The PoC validated end-to-end data flow, context reasoning, and natural language interaction on an Android prototype.

See our Presentation materials

AI Minutes, an AI-powered transcription service

AI Minutes is designed to securely and accurately generate meeting minutes for various corporate settings. As the Product Manager, I lead our AI and Development teams in integrating advanced technologies into our web application. This includes developing and implementing Speech-to-Text models, Speaker Diarization models, and Large Language Models to enhance transcription accuracy and speaker identification.

Check our Product

A Multi-Target Feature Selection Method in Predicting Psychological Status with Text Data

Based on previous literature on Multi-Target Feature Selection (MTFS) and psychology, this research proposes a MTFS method to mitigate overfitting when predicting psychological status using text data.

Paper published in Information Integration and Web Intelligence (iiwas 2024)

Academic poster presenting NLP analysis of expert feedback for IPBES regional assessments with multiple text sections and charts at Tsukuba Conference 2023 in a formal conference setting

Utilizing NLP techniques to Analyze Expert Feedback on IPBES's Regional Assessments

The IPBES is an intergovernmental body that has taken significant role in examining biodiversity trends. In this project, we leverage NLP techniques to undertake contextual semantic network examination of the expert comments from IPBES.

Poster accepted by Tsukuba Conference 2023

Academic poster titled University Students Temporal Trends on X and Their Effects on Subjective Well-being during the COVID-19 Pandemic with graphs and explanatory text displayed in a calm research environment

University Students' Temporal Trends on X and Their Effects on Subjective Well-being during the COVID-19 Pandemic: Examination from a Two-Year Study of Posts

Utilizing NLP techniques, we extracted nouns from Japanese tweets as language features. These features were then used in regression analyses to link with subjective well-being scores, followed by visualization through a new word cloud technique.

Paper accepted by DEIM 2024

Lyrics Analysis on Songs made by John Mayer

I analyze the lyrics of songs made by John Mayer using NLP.

Link to the full project

Blogs

Utilizing NLP to Analyze Expert Feedback on IPBES's Regional Assessment

I walk through how qualitative expert feedback can be transformed into structured insights using topic modeling, sentiment analysis, and clustering to support biodiversity research. Read More

BERT vs. LSTM: Performance in Sentiment Classification

I compare BERT and LSTM models on accuracy and training time for sentiment analysis, showing where transformer-based approaches shine and where lighter recurrent models still win. Read More

Loading latest Medium posts…

Education

MS in Informatics of the Graduate School of Comprehensive Human Sciences

University of Tsukuba, 2022-2024

Certificate Programme on Nature Conservation

University of Tsukuba, 2022-2024

MA in Area Studies of the Graduate School of Humanities and Social Sciences

University of Tsukuba, 2018-2021

Exchange Program (Coursework)

Washington University in St. Louis, 2019-2020

BA in International Relations of the School of Social and International Studies

University of Tsukuba, 2014-2018

Experiences

AI Engineer at FCNT (Lenovo Group)

Years Worked: May 2025 - Current

I lead the Japanese testing team, ensuring that AI features are effectively adapted to local contexts and user needs. I provide weekly progress reports on AI localization to teams in China, the US, and India, enabling alignment across regions. I also work closely with Brazilian and Chinese teams to refine prompts, resulting in more accurate and user-friendly AI outputs.

Product Manager in AI Application Development at Commerce Robotics Inc.

Years Worked: Oct. 2024 - Apr. 2025

I manage a team of AI engineers in Vietnam and developers in India, collaborating to create innovative AI-driven applications. Additionally, I contribute to the Global Division Department by driving strategies to enhance sales performance in Japan's online shopping marketplaces.

LLM Intern at Free Web Hope Inc.

Years Worked: Feb. 2024 - Sep. 2024

I collaborated on the LLM-Creative project, focusing on understanding the company's workflow and exploring improvements to the creative production process using LLMs for more efficient, reproducible outputs. Additionally, I researched how to use LLM tools for generating advertising images through Midjourney and explored video generation methods like Runway Gen-3. I also developed a customized ChatGPT model that generates catchy copywriting from URLs and utilized OpenAI's API in Google Apps Script to automate customer and market data analysis, enhancing workflow efficiency.

Research Intern at Complexity Science Hub

Years Worked: July 2024 - Aug. 2024

I focused on network analysis of conspiracy groups on Telegram, creating bipartite networks to explore the interactions between groups and authors. I utilized tools like NetworkX and Gephi to calculate key metrics such as degree centrality, betweenness centrality, and PageRank. I categorized groups based on the number of authors and messages, analyzed shared authors between groups, and developed methods to calculate group closeness.

Intern at International Union for Conservation of Nature Japan Committee (IUCN-J)

Years Worked: Aug. 2024 - Sep. 2024 & Sep. 2023 - Nov. 2023

During my internship at IUCN-J, I participated in The 25th Subsidiary Body on Scientific, Technical and Technological Advice (SBSTTA 25) in Nairobi, Kenya from 15-19 Oct 2023, as well as the 8th IUCN Asia Regional Conservation Forum (RCF) in Bangkok, Thailand from 3-5 Sep 2024. My role involved writing reports in English and Japanese to summarize key points from the conference, conducting interviews with delegates, and writing articles based on those interviews.

Writing Samples: Reimagining Conservation in Thailand: A Nature Positive Future, Bridging Perspectives: Climate and Biodiversity Conversations from Egypt and Somalia

Technical Assistant in Dabot Chatbot Project at University of Tsukuba

Years Worked: May 2023 - July 2024

"Dabot: Mr. Dax's Consultation Room" is a chatbot that automatically suggests useful coping information for people with a variety of neurological characteristics, mainly in the college age group. In this project, my responsibilities involved writing Google Apps Scripts for task automation and visualizing user data using Python.

Account Manager of Chinese Social Media at Yamada Bee Farm

Years Worked: April 2021 - July 2022

In this role, I oversaw official accounts across diverse social media platforms and fostered collaborations with business partners in Japan, mainland China, Taiwan, Vietnam, and Singapore. This experience notably contributed to the development of my adept communication and negotiation skills. It was during this period that my interest in both social media research and the practical applications of natural language processing in business was sparked.

Contact

If you'd like to get in touch, feel free to email me at: caidanmeng@gmail.com