✧Auto-evaluation system ● Led the initiative to promote Test-Driven ● Development methodologies across multiple teams, initiating the design and implementation of an auto-evaluation system architecture. ● Led the transformation from a paid model into an open-source structure, enhancing access to top-tier resume tools. Created an LLM auto-generative chatbot system for recommendation assessments, search functions, and interactions, overseeing team management to ensure scalability and user-centric alignment. ● Bring extensive experience in data-driven analysis, leveraging insights for strategic decision-making and optimization. Proficient in data visualization and statistical analysis, adept at transforming complex datasets into actionable insights. ● Contributed through collaborative leadership, employing data analysis to offer strategic guidance and optimization solutions. Focused on delivering impactful, technology-driven solutions that add value and improvements. ✧Auto-tagging recommended system ● Led the design and implementation of an architecture that reduced the need for 25 annotators, enhancing the speed by 240 times for video review content tagging and filtering API. ● Designed and optimized video ML pipelines, significantly improving moderation video content accuracy from 60% to approximately 80% and scaling up object detection items by over 10 times. ● Designed and implemented data pipelines to increase speed 20x by using SNS, SQS and lambda technologies on AWS to solve traffic problems and monitoring of API operations using AWS CloudWatch. ● Developed and implemented storage transformation pipelines from Google storage to AWS S3 and back to GCP cloud storage, streamlining storage transformation and lifecycle management. ● Utilized BigQuery and Spanner to analyze historical top tags, accounting for 99% of usage, for tag situation assessment and user behavior analysis, facilitating informed label decisions. ● Created an image quality detection system for videos, establishing an automated and objective screening mechanism to replace human annotators. ✧Other automations ● Enhanced accuracy from 50% to 90% using Graph theory methods to expand test coverage and improve model performance within the workflow generation. ● Implemented innovative embedding and transformer methods for information extraction while developing a vector DB on Qdrant and similarity search API to enhance efficient data retrieval.

✧ Forecast future retail sales and optimize purchase, sales and inventory ● Designed and implemented ETL for feature engineering for prediction models by crawling weather websites, and using feature engineering to create more than 50 features. ● Designed and compared models performance for multiple forecasting tasks to evaluate a final deployment model. ● Designed and optimized store stock by using forecasting model and business consideration. ✧ Others ● Designed and implemented features for prediction models by crawling websites. ● Designed and implemented a forecasting model that predicted trends in product outcomes.

✧ Wave-CAIPI MPRAGE pediatric testing project (Collaborate with Siemens) ● Led the test of pre-production sequence research to reduce the scan time from 5.5 minutes to 3.5 minutes on pediatric image of WAVE-CAIPI MPRAGE in clinical use. ● Designed and implemented the development of an A/B test matrix for the evaluation of pediatric brain between different imaging sequences data. ✧ Mild traumatic brain injury translational research study ● Designed and implemented an analytic method to discover imaging biomarkers. ● Designed and implemented longitudinal functional MRI and diffusion kurtosis imaging data, employing ICA and logistic modelsto make future predictions for future diagnosis insight, and authored results for publication. ✧ Taipei Medical University Lung cancer label database ● Designed and led the optimization for over 5000 semantic and CT lung cancer labeling. ✧ Diffusion kurtosis imaging on spinal cord condition diagnosis ● Designed and implemented the analysis of diffusion kurtosis imaging on spinal cord to gain new type condition diagnosis insight.✧ Wave-CAIPI MPRAGE pediatric testing project (Collaborate with Siemens) ● Led the test of pre-production sequence research to reduce the scan time from 5.5 minutes to 3.5 minutes on pediatric image of WAVE-CAIPI MPRAGE in clinical use. ● Designed and implemented the development of an A/B test matrix for the evaluation of pediatric brain between different imaging sequences data. ✧ Mild traumatic brain injury translational research study ● Designed and implemented an analytic method to discover imaging biomarkers. ● Designed and implemented longitudinal functional MRI and diffusion kurtosis imaging data, employing ICA and logistic modelsto make future predictions for future diagnosis insight, and authored results for publication. ✧ Taipei Medical University Lung cancer label database ● Designed and led the optimization for over 5000 semantic and CT lung cancer labeling. ✧ Diffusion kurtosis imaging on spinal cord condition diagnosis ● Designed and implemented the analysis of diffusion kurtosis imaging on spinal cord to gain new type condition diagnosis insight. ✧ Paper： Sho-Jen Cheng, Ping-Hui Tsai, Yun-Ting Lee, Yi-Tien Lee, Hsiao-Wen Chung *, Cheng-Yu Chen. (in press). Diffusion tensor imaging of the spinal cord. Magnetic Resonance Imaging Clinics of North America. ✧ Conference paper： Yun-Ting Lee, Chia-Feng Lu, Nai-Chi Chen, Li-Chun Hsieh, Sho-Jen Cheng, Yu-Chieh Jill Kao, and Cheng-Yu Chen. (2019) Anxiety-Related Alterations of Resting-State Networks in Mild Traumatic Brain Injury. Poster presented at the 49th Society for Neuroscience annual meeting. October 19-23 2019, at McCormic Place, Chicago, Illinois, America.

✧ 職責內容: 1. 利用人類行為資料及fMRI資料以統計方法分析，找出網路成癮的影像　 　重要特徵，及製作結構方程人類網路成癮互動之行為模型。 2. 以DTI影像以統計方法分析，找出老化與特定認知功能相關之重要特徵。 程式語言：MATLAB、Linux 作業系統/工具：Windows, Linux ✧ 專案或產品計畫: 1. 老化大腦連結與認知控制:評估與介入橫斷後續研究 2. 科技部計畫:網路成癮傾向程度對自我與他人互動作業的影響 3. 科技部計畫:數位科技與多重作業表現 ✧ Paper： Chang, Y.-H.*, Lee, Y.-T., Hsieh, S.* (2019). Internet Interpersonal Connection Mediates the Association between Personality and Internet Addiction. International Journal of Environmental Research and Public Health, 16, 3537. ✧ Conference paaper： Yun-Ting Lee, Shu-Lan Hsieh. (2015) The correlation between Feature of Neuroticism and Online interpersonal feelings with internet addiction. Poster presented at the 55th Taiwanese Psychological Association Annual Convention and International Psychological Conference, October 15-16 2016, at NCKU, Tainan, Taiwan. Thesis： Yun-Ting Lee. (2018) The Influence of Feelings Internet interaction and Neuroticism on Internet Addiction. National Chen Kung University.

Web design, maintain