User details
Wei Hong Low
@wei_hong_low_wrz4pdd6p
Data Scientist
Singapore, Singapore
2Followers0Following0Mutual0Post
Member since: Jan 21, 2025I am a data scientist who had 5 years in depth experience in NLP modellling and deployment.
• Constructed a centralized web crawling platform, Crawlab, from scratch, overseeing the operation of over 40 active crawlers.
• Engineered a comprehensive monitoring dashboard, leveraging data from web crawlers on platforms such as YouTube, Twitter,
and various crypto exchanges which enabled the business development team to effectively track and analyze the activities of key
opinion leaders (KOLs) in the crypto space.
• Automated the entire pipeline for identifying new leads from YouTube, specifically targeting crypto KOLs, and distributing
these leads to business development teams across various countries.
• Fine-tuned and deployed over 13 translation models, achieving a significant cost reduction of 35% of existing vendor costs.
• Developed an automated intelligence news distribution system, achieving an accuracy rate of over 95% in distributing news to
relevant groups.
• Supported credit risk team to crawled more than 200k Whatsapp profile images.
• Supported teammates with crawlers such as Taobao, Tianmao, JingDong, Smartprix, Gadgetsnow, dpreview, devicespecs and
Instagram.
• Shared crawling techniques during a team internal meeting.
• Person In Charge of Keyword Spam Models for all markets who trained and maintained the precision Logistic Regression models
combined with feature extraction technique over 6 markets regionally over 98%.
• Built and maintained the accuracy of Standard Product Unit (SPU) models over 6 markets regionally over 97% for Laptop.
• Built and maintained all the 7 daily job pipelines using airflow.
• Involved in keyword mining projects to mine representative keywords for each category.
• Trained machine learning models such as Fasttext and ElectraBert to classify items to its correct category with accuracy over
90% for all categories in Indonesia market.
• Implemented text mining technique to mine important product keywords for each level 3 category.