Cake Talent Search

麻將AI的研究開發 參加Computer Olympiad麻將程式競賽得過4次金牌 熟悉Monte Carlo Tree Search, Important Sampling等蒙地卡羅類型 演算法 熟讀數篇AlphaZero, MuZero類型演算法之期刊論文、國際會議論文 了解常見深度學習網路架構 Transformer, Resnet等 接觸過深度強化式學習演算法 DQN, A3C, DDPG, PPO等 Imitation learning, inverse

學習經歷 東捷資訊股份有限公司AI研發工程師 工作內容：AI 相關應用與模型訓練 影像辨識：人臉辨識、OCR、車牌辨識 最佳化演算法：FJSP、裝箱、路順模擬、先進排程 專案管理 新人類BPM電子簽核系統專案管理以及開發 瑞台電傳股份有限公司軟體工程

功能或優化方法，促進經驗交流。 AUO 友達光電 • 八月五月 2023 工廠廢水處理流程節能專案 : 主要使用Python OpenCV及C#，開發影像辨識演算法，將影像中特定數值量化並 結合多種感測器的數值， 結合最佳化演算法，以最佳參數控制機台，達到節能及降低化學

服專案，與客服部門協作，在六個月內成功上線功能並持續迭代優化。 5. 與醫美事業體客服部門密切合作，根據需求開發演算法，解決客服約診操作上不易安排客人預約的痛點，使客戶約診數提升20%。 6. 從客服與客戶對話紀錄中，察覺用戶對

accomplished by Elo score calculation and optimizing the recommendation system for new user card stacks, prioritizing user-centric engagement. Enhanced the Top Picks Recommendation system, achieving a 200% increase in click rate and generating 30% of revenue from top picks by implementing an Elo ranking system. Rewrote the algorithm using Cython, resulting in a 300% reduction in calculation time. Improve paid conversion rates by 300%. Identified crucial insights from the behavior of D0 users with key factors, and developed models for D1-D7 by using XGBoost, achieving 85% hit rate in the top 5% of users

. WORK EXPERIENCE Data Scientist • Wistron FebNow Develop deep learning models to solve the problem of computer vision, generative AI and LLM Develop machine learning models to do numerical analysis Design APIs and deploy the services via virtual environment Dig out beneficial information from complex data to support decision-making Algorithm Engineer • Foresight Technology OctFebEnhance the algorithm of Automatic Virtual Metrology System(AVM) Apply AVM in production system Industrial Engineer • Winbond MarAugAnalyze the manufacturing data to improve the bottleneck of process Develop real-time reports for monitoring the production status EDUCATION National Cheng Kung University M.S. Industrial

Design and architect innovative AI product systems - Mentor team members and provide expert AI technical guidance - Spearhead the development of major AI products: Key Projects: 1. Travel Plan Recommendation System - Integrated LLMs, RAG, and knowledge graphs for personalized travel planning - Implemented route optimization using GEO libraries and shortest path algorithms - Developed feedback loop for continuous model improvement using Neo4j and FAISS - Technologies: LLM, RAG, Neo4j, FAISS, GEO libraries, PyTorch, PEFT, LORA 2. CreativeMAX - Marketing Image Evaluation System - Consulted on AI-driven image analysis for marketing effectiveness - Guided implementation of eye-tracking prediction and artistic evaluation - Technologies: Computer Vision