Tsai Chen Hsiung
MSCS @UTD | Data Scientist @ SparkAmplify
Hire this person
Send a job offer directly to this candidate
About
Enthusiastic, fast-learning data scientist with 4 years hands-on experience in ML and excellent communication skills seeking for a challenging internship opportunity in a progressive company to utilize modeling and analytics skills to create value.
Experience
● Built up data pipeline integrating web crawling with MySQL database with task-tracking push notification.
● Collected over 3000 news by labor saving data crawling pipeline that integrated AWS Lambda and Zyte.
● Auto-processed 300k articles hourly with PageRank keyword model for convenient access to a 10 members team.
● Clustered industry news to extract trending events using unsupervised learning methods: LDA, HDBSCAN.
Projects
- Seasonal industrial trending topics, AI Hub, Taiwan.
₋ Reached human-picked accuracy of trending market topics through XGBoost text classifier.
₋ Saved 70% of customer time discovering the stakeholder influencers by deploying SNA analysis.
- Potential LinkedIn customer matching, Woodmax Taiwan & Conquer Co., Ltd.
₋ Utilized AWS and cloud functions to automate data scraping and analysis pipeline.
₋ Cross-matched 10000 profiles within 10 seconds by auto-summarizer to rank top potential customers.
Education
Selected courses
Machine learning, Database design, Advanced statistics in data science, Data Structure and algorithm
● Led five undergraduates student team to build NLP-based Youtube video summarizer as Chrome extension.
● Achieved <3 sec video summary by integrated AWS service: S3, API Gateway, Sagemaker and HuggingFace model.
● Implemented error-shooting ensemble model resulting in 15% improvement on user satisfaction.
Projects
Athena, Axxess Hackathon Winner
● Implemented Firebase and Flask base backend server to support API requests interaction.
● Augmented user query information by 25% using Roberta sentiment analysis to detect users’ prompt emotions.
● Integrated ChatGPT into voice-to-text Whispir service to reach over 90% satisfaction in judges’ selection
Next Big Coin, HackDFW Top5
● Created a monitoring web app to help the next generation with their financial management.
● Designed economic indexes to inspect users’ finance based on their investments, expenses, and income.
