AI consultant 2021/9 - Current Deloitte
- Develop computer vision model to monitor animal welfare in slaughterhouse
Who Am I?
Hi I'm Kira Liang I am a data scientist from Taiwan, currently living in the Netherlands. I have a passion for data and several hands-on experiences in solving business problems across multiple domains. During my academic training and working experience, I have been a specialist at managing complex raw datasets and transferred them into valuable insights to achieve objectives. Also I enjoy working in a team, I believe you can learn faster and much more.
Creative
Detail-oriented
Team work
Initiative
Cities
Countries
Projects
Languages
These are some of my expertise
Data Analytics
Data Mining
Machine Learning
Natural Language Processing
Information Retrieval
Big Data
Python
SQL
R
HTML,CSS
AWS
JAVA
Education
University of Amsterdam
2018/9 - 2019/8
The Netherlands
- Translate scientific, societal and business challenges into data related solutions
- Machine learning and deep learning techniques
- Big data cloud application
Institute for Information Industry
2016/12 - 2017/6
Taiwan
- Data analytics application:
- Programming in Java, R, python
- Big Data framework(Hadoop ecosystem)
- Relational and non-relational database applications (SQL server & MongoDB)
National ChengChi University
2012/9 - 2016/6
Taiwan
- Micro/macro economics
- Statistics
- Financial management
Work Experience
Data analyst 2020/3 - 2021/7 Zivver
- Create ELT data pipeline for commercial data
- Conduct statistical modeling to facilitate business decisions
- Provide actionable insights, recommendations to the sales and marketing teams.
Machine learning specialist intern 2019/4 - 2019/6 Elsevier
- Developed a deep learning model (BERT: Deep Bidirectional Transformers) that utilized text features from raw noisy abstract and title to correctly classify multi-labeled scientific articles
- Outperformed the baseline model at 39% with respect to accuracy
Data Scientist 2018/4 - 2018/8 Digiwin Software Co., Ltd.
- Developed insightful metrics, dashboards, reports and insights for the executive team
- Prediction of retirement: Quantified retirement reasons, reduced employee turnover rate by 12% and reduced business operation cost by 22%
- Product sales forecast: Gauged product demand and increased inventory turnover rate by 15%
Data Scientist 2017/4 - 2018/3 ProphetStor Data Services, Inc
- Developed conversational flows for chatbots
- Created data pipelining to using the chat conversations to train the AI Bot
Projects
| NLP, Machine learning, Deep learning |
Multi-label text classification with BERT
Text classification is a common task in Natural Languge Processing (NLP), in this research I compared BERT, a popular Transformer model and other classical models on scientific contents.
| Analytics, prediction modeling |
Theft crime prediction based on spatial and temporal crime hotpots
This study aims to find the spatial and temporal theft hotspots in Taipei city by analyzing real- world theft datasets through a statistical analysis and the utility of hotspot mapping.
| NLP, Machine learning, Deep learning |
Identify duplicated question pairs for Quora
Can we identify the intent of a question even if they are phrase differently? In this project, we trained more than 300k question pairs from quora and identify if they are duplicated.
