Yi Cyuan Lin Email & Phone Number

Taipei, Taipei City, Taiwan

Taiwan

• As a contributor – Built an AD auto-strategy system which improved 120% CTR and reduce 23% CPC. – Built an invalid-traffic recognition system which avoided millions of AD request calculations. – Built an AD creative production base on GenAI(stable diffusion) and share the result on 2023 AWS TW Summit. – Developed an AD CTR algorithm which improved 80% CTR performance on offline test. – Built a Brand Left Survey report which help the advertiser to evaluate the AD performance with questionnaire. – Built a system to predict the AD traffic with the strategy setting and provide a suggestion. It reduce 50% time for the Media Develop team. – Built a system to provide a maximum impress plan of DOOH AD Campaign. – Built contextual targeting and cross screen audience targeting advertisement system. – Planed a cookieless solution.• As a team leader – Import Airflow and AWS Sagemaker service to reduce 20-80% cost. – Led four data scientist to develop real time bidding, user base and content base AD CTR model and user profile labeling. Consequently, AD CTR performance arise 20-50%. – Led a data analysis to analyze advertisements’ insight of festival and our AD network.

台灣 Taipei City 台北

• Built an also like algorithm which extent 20-50% number of audience with the same AD CTR performance.• Developed an AD CVR algorithm on AD network. It improve AD 100% CTR performance.• Built an audience report system which include AD network, telecom, and invoice data analyze and visualize the result on looker studio. It got 80% satisfaction from the sales team.• Developed an audience score metric which helped us to evaluate websites’ value.• Optimize crawler system and reduce 30% of the cost.• Built a first party cookie system.

台灣 臺北市

• Built a blog articles’ recommendation algorithm on web push and improve the CTR 100%. In this system, there are millions users and articles.• From end-to-end built an blog articles’ recommendation system on the bottom of the page. As a result, the CTR 10% better than the competitor.• Developed an e-commerce CVR algorithm on web push.• Improve user behavior collection system from a VM to an auto-scaling system within 10 second latency.• From end to end build a video advertisement system. Using FFMPEG to transform video as m3u8 format reduce 30% cost of traffic fee.• Built a line chatbot system which can automatic reply for specific questions and send message to slack. It helped the sales team arise 30% work performance.

碩士, Information Technology, Gpa: 3.8 / 4.0

Thesis title: Reduction in Number of Constraints in Max-Margin Early Event Detectors. Journal: Pattern Recognition Volume 91, July 2019.

Intensive English Program, Level 700

Finish the highest level English Program in UCR.

學士, Information Technology

Activities and Societies: Guitar club and Pop Music clubProduced Graduation song : https://www.youtube.com/watch?v=uUmRulY8OFI