Zhe-Li Lin (林
林
林哲
哲
哲立
立
立)
joli79122@gmail.com. (886)925680266 Github:https://github.com/JoliLin
PROFILE
Data Science researcher with 6 years of experience in NLP, recommendation systems, social networks, machine
learning, and deep learning modeling. Published an NLP paper at the PAKDD conference with source code on
GitHub. Passionate about researching and implementing state-of-the-art algorithms and models. Additionally, I
have over a year of experience as a machine learning engineer, where I led projects, conducted analyses, and oversaw
the development of systems and product management.
WORK EXPERIENCE
• Machine learning engineer Jun. 2022 - Present
Rosetta.ai
– Built a Recommender System with about 33% performance improvement and reduced computing cost by
28% compared to the previous system
The framework of the system is combined with a embedding learning part which contains Deepwalk,
LINE and GraphSAGE and a classifier part which contains FM, cosine-similarity and dot product
– Dockerized the recommendation system and deploy the system through Amazon beanstalk
– Cleaned the raw data of customers’ behavior in mariadb and using Amazon Redshift to save the clear
data
– Built a data pipeline to analyze the performance of customers and using redash to visualize the result
– Constructed a search engine through Amazon Opensearch for searching a product from our customers’
products
– Fetched the features by using Google Product API such as material, style of customers’ products
– Used Microservice Architecture to construct the system and built API of each function by FastAPI
– Built a chatbot by using OpenAI API and using gradio to build the front-end, also compared the
performance with other 5 open sources LLM such as: RWKV, Cerebras-GPT, Alpaca, ChatGLM
and OpenChatKit
– Scheduled the tasks such as training recommendation system and cleaning data by using airflow
– Managed the servers with Ubuntu system on Amazon AWS
• Personal studio Feb. 2021 - Jun. 2022
– Constructed a Chinese search engine and a questionnaire recommendation system built by Elasticsearch
and Django for prototype website
– The questionnaire recommendation system model is trained by BERT which is using BERT to compute
the embedding of each questionnaire and compute cosine similarity to recommend similar questionnaire
• Research Technician Assistant June. 2020 - Jan. 2021
Communication Data and Network Analytics lab, Center for Survey Research, Academia Sinica
– Managed the servers with Ubuntu system
– Served as a data scientist and a backend engineer also as a team leader
– Built a automatic crawling system to crawled the data from 79 News websites in Taiwan containing
Facebook, Twitter, Dcard, ptt and Line
– Built a search engine based on the crawled data via Elasticsearch with Flask
– Trained a political polarity predicting model using CNN model and built a prototype website with Flask
• Research Assistant Jan. 2017 - Jan. 2020
Computational Linguistics & Information Processing Lab., Academia Sinica
– Handcrafted the deep models and NLP algorithms: CNN, RNN, LSTM, bag-of-word model, word2vec
and BERT
– Published the Character-level Convolutional Neural Tensor Networks model and research
multi-task model and proposed Disposable Auxiliary Network
– Published papers on journal Social Network Analysis and Mining and conference WebSci’15,
ACSS’16, PAKDD’19 and WI-IAT.
1
PUBLICATIONS
• Journal Papers
– Ming-Feng Tsai, Chih-Wei Tzeng, Zhe-Li Lin, and Arbee L.P. Chen. Discovering Leaders from Social
Network by Action Cascade. Social Network Analysis and Mining, 4(1): 165, 2014.
• Conference Papers
– Ming-Feng Tsai, Chuan-Ju Wang, and Zhe-Li Lin. Social Influencer Analysis with Factorization Ma-
chines. In Proceedings of the 2015 ACM Web Science Conference (WebSci ’15), Oxford, 2015.
– Zhe-Li Lin, Yu-Ming Lu, Ming-Feng Tsai, and Chuan-Ju Wang. Measuring Social Influence on Online
Collaborative Communities. In Proceedings of the 7th Asian Conference on Social Sciences (ACSS ’16),
Kobe, 2016.
– Zhe-Li Lin and Chuan-Ju Wang. Keyword Extraction with Character-level Convolutional Neural Tensor
Networks. In The Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD ’19).
https://github.com/cnclabs/CNTN
– Chih-Ting Yeh, Zhe-Li Lin, Sheng-Chieh Lin, Jing-Kai Lou, Ming-Feng Tsai, and Chuan-Ju Wang, ”A
Learning Framework with Disposable Auxiliary Networks for Early Prediction of Product Success” In the
20th IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology
(WI-IAT)
EDUCATION
• National Chengchi University Taipei, Taiwan
Master in Computer Science Sep. 2012 - July 2016
• University of Taipei Taipei, Taiwan
Bachelor in Computer Science Sep. 2008 - June 2012
SKILLS
• Languages and Systems: Python, Java, C, Javascript, bash, MS, Unix-like systems
• Deep Learning Framework: Pytorch, Chainer, Tensorflow, Keras
• API and Web Tools: Django, Flask, FastAPI
• Cloud Platform: Google Cloud Platform, Amazon AWS(EC2, Redshift, Beanstalk, Opensearch)
• Database: MariaDB, MySQL, Redis, Elasticsearch
• Search Engine Framework: Elasticsearch
• Tools: Git, Docker, Airflow, LaTeX
2

CV.pdf

  • 1.
    Zhe-Li Lin (林 林 林哲 哲 哲立 立 立) joli79122@gmail.com.(886)925680266 Github:https://github.com/JoliLin PROFILE Data Science researcher with 6 years of experience in NLP, recommendation systems, social networks, machine learning, and deep learning modeling. Published an NLP paper at the PAKDD conference with source code on GitHub. Passionate about researching and implementing state-of-the-art algorithms and models. Additionally, I have over a year of experience as a machine learning engineer, where I led projects, conducted analyses, and oversaw the development of systems and product management. WORK EXPERIENCE • Machine learning engineer Jun. 2022 - Present Rosetta.ai – Built a Recommender System with about 33% performance improvement and reduced computing cost by 28% compared to the previous system The framework of the system is combined with a embedding learning part which contains Deepwalk, LINE and GraphSAGE and a classifier part which contains FM, cosine-similarity and dot product – Dockerized the recommendation system and deploy the system through Amazon beanstalk – Cleaned the raw data of customers’ behavior in mariadb and using Amazon Redshift to save the clear data – Built a data pipeline to analyze the performance of customers and using redash to visualize the result – Constructed a search engine through Amazon Opensearch for searching a product from our customers’ products – Fetched the features by using Google Product API such as material, style of customers’ products – Used Microservice Architecture to construct the system and built API of each function by FastAPI – Built a chatbot by using OpenAI API and using gradio to build the front-end, also compared the performance with other 5 open sources LLM such as: RWKV, Cerebras-GPT, Alpaca, ChatGLM and OpenChatKit – Scheduled the tasks such as training recommendation system and cleaning data by using airflow – Managed the servers with Ubuntu system on Amazon AWS • Personal studio Feb. 2021 - Jun. 2022 – Constructed a Chinese search engine and a questionnaire recommendation system built by Elasticsearch and Django for prototype website – The questionnaire recommendation system model is trained by BERT which is using BERT to compute the embedding of each questionnaire and compute cosine similarity to recommend similar questionnaire • Research Technician Assistant June. 2020 - Jan. 2021 Communication Data and Network Analytics lab, Center for Survey Research, Academia Sinica – Managed the servers with Ubuntu system – Served as a data scientist and a backend engineer also as a team leader – Built a automatic crawling system to crawled the data from 79 News websites in Taiwan containing Facebook, Twitter, Dcard, ptt and Line – Built a search engine based on the crawled data via Elasticsearch with Flask – Trained a political polarity predicting model using CNN model and built a prototype website with Flask • Research Assistant Jan. 2017 - Jan. 2020 Computational Linguistics & Information Processing Lab., Academia Sinica – Handcrafted the deep models and NLP algorithms: CNN, RNN, LSTM, bag-of-word model, word2vec and BERT – Published the Character-level Convolutional Neural Tensor Networks model and research multi-task model and proposed Disposable Auxiliary Network – Published papers on journal Social Network Analysis and Mining and conference WebSci’15, ACSS’16, PAKDD’19 and WI-IAT. 1
  • 2.
    PUBLICATIONS • Journal Papers –Ming-Feng Tsai, Chih-Wei Tzeng, Zhe-Li Lin, and Arbee L.P. Chen. Discovering Leaders from Social Network by Action Cascade. Social Network Analysis and Mining, 4(1): 165, 2014. • Conference Papers – Ming-Feng Tsai, Chuan-Ju Wang, and Zhe-Li Lin. Social Influencer Analysis with Factorization Ma- chines. In Proceedings of the 2015 ACM Web Science Conference (WebSci ’15), Oxford, 2015. – Zhe-Li Lin, Yu-Ming Lu, Ming-Feng Tsai, and Chuan-Ju Wang. Measuring Social Influence on Online Collaborative Communities. In Proceedings of the 7th Asian Conference on Social Sciences (ACSS ’16), Kobe, 2016. – Zhe-Li Lin and Chuan-Ju Wang. Keyword Extraction with Character-level Convolutional Neural Tensor Networks. In The Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD ’19). https://github.com/cnclabs/CNTN – Chih-Ting Yeh, Zhe-Li Lin, Sheng-Chieh Lin, Jing-Kai Lou, Ming-Feng Tsai, and Chuan-Ju Wang, ”A Learning Framework with Disposable Auxiliary Networks for Early Prediction of Product Success” In the 20th IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) EDUCATION • National Chengchi University Taipei, Taiwan Master in Computer Science Sep. 2012 - July 2016 • University of Taipei Taipei, Taiwan Bachelor in Computer Science Sep. 2008 - June 2012 SKILLS • Languages and Systems: Python, Java, C, Javascript, bash, MS, Unix-like systems • Deep Learning Framework: Pytorch, Chainer, Tensorflow, Keras • API and Web Tools: Django, Flask, FastAPI • Cloud Platform: Google Cloud Platform, Amazon AWS(EC2, Redshift, Beanstalk, Opensearch) • Database: MariaDB, MySQL, Redis, Elasticsearch • Search Engine Framework: Elasticsearch • Tools: Git, Docker, Airflow, LaTeX 2