林奕勳
AI Engineer
- 2+ years of financial AI development experience, specialized in the design of python API, SDK, and framework for optimizing the deployment flow of data science, ETL, and machine learning pipelines.
- 6+ years of deep learning research and development experience, with a strong ability in prototyping, evaluation, and deployment of models, in both academical and industrial environment.
- 1+ year of participation in the technology management assistant bootcamp, with real-experience on business user support, gaining first-hand insight on financial and corporative AI usage via cross-department collaboration.
- Having the quality of perseverance in face of challenging problems and keep absorbing “technical-nutrients” from the development of python open-source community.
玉山銀行智能金融處 (Intelligent Finance Dept., E-SUN Bank )National Taiwan University
Taipei, Taiwan
Professional Background
Current status
Employed
Profession
Data Engineer・Machine Learning Engineer
Fields
Software・Banking
Work experience
1-2 years (4-6 years relevant)
Management
I've had experience in managing 1-5 people
Skills
ResearchMusic Information RetrievalNatural Language ProcessingInformation RetrievalMachine Learningpython programmingC++ LanguageJavaTensorflow (Keras)Numpy Data ManipulationsklearnLinuxVerilogArduinoGithubAirflowDesign PatternsDockerPython ProgrammingPyTorchPytorch LightningPostgreSQL
Languages
English・Intermediate
Highest level of education
Master
Job search preferences
Desired job type
Full time・Interested in working remotely
Desired positions
Machine Learning Researcher / Engineer
Desired work locations
台灣台北市・台灣新竹市
Freelance
Part-time freelancer
Work Experience
人工智慧工程師
玉山銀行智能金融處 (Intelligent Finance Dept., E-SUN Bank )・Full-Time
August 2020 - Present2 yrs7 mos
台灣台北
ETL框架設計：主導設計結構化資料框架以幫助公司資料科學團隊在低轉換成本將類pandas資料處理pipeline導入airflow自動排程系統，並同時提供底層自動銜接CuDF加速容器功能，使資料運行瓶頸能達到400倍以上加速。 ETL效能提升研究：指導實習生建置類pandas套件的速度與空間效能的profiling實驗框架與模擬環境，以協助資料科學團隊快速評估各種ETL元操作與其使用之套件的最佳組合方案。 信貸行銷方案產製專案：在此跨部門合作專案中，代表AI部門協助與指導業務部門以python與SQL將業務邏輯整合進框架中，主導行銷方案運算的框架的設計以支援未來貸款方案擴增彈性及跨部門協作的方便性，另外，也針對運算瓶頸協助透過多進程與多容器運算將700萬顧客資料的inference 時間從5天加速到8小時。 房貸比較法Graph 模型研發：深入了解不動產估價技術法規，改寫學界的深度學習房屋估價模型以加入符合法規要求的解釋性，並以Google SCANN加速估價案例發查速度，以提升模型training 與inference data的產製速度以及提升可擴充的彈性。 Deep Learning 實驗框架設計：以pytorch-lightning 與lazy-computation 為核心建構規格化的深度學習實驗框架，以加速外部合作學術機構所研發的模型架構導入，並以Ray tune 建置多進程平行調參功能以加速模型優化速度。 GitHub SideProjects: - JSON Schema 自動生成 / Python DataClass 自動生成 - 資料管線 Monad Design Pattern 研發 - PyPi 套件Dependency Graph爬蟲 - 加密貨幣市場趨勢分析平台API - Python Multiprocessing 平行運算工具研究 - Ray Multiprocessing Actor 實作開發 - Python AST 語法樹研究 - 基於Ray的模型自動調參工具研究產學合作研究生
願境網訊股份有限公司 KKBOX Taiwan Co., Ltd.
September 2015 - January 20204 yrs5 mos
Research and develop machine learning techniques for music recommendation and retrieval system under programs between MPAC lab and KKBOX. Programs: [1] https://www.grb.gov.tw/search/planDetail?id=12487954 [2] https://www.grb.gov.tw/search/planDetail?id=12375507 Publication: https://www.eurasip.org/Proceedings/Eusipco/Eusipco2018/papers/1570437289.pdf Research Summary: https://www.evernote.com/shard/s89/nl/19207065/d165721b-0951-4e23-a783-9a4382ce6ebe Commends of my journal paper: https://www.evernote.com/l/AFk1iCBz5UhOLJowRl3klttz8BEmW38gfBI/App Software Development Intern
July 2014 - August 20142 mos
Analyzed app UI design and proposed an improved designed
Education
National Taiwan University
Master’s Degree・Graduate Institute of Communication Engineering
2015 - 2020
National Tsin-Hua University
Bachelor’s Degree・Department of Electrical Engineering
2011 - 2015・ 3.9/3.9 GPA
Activities and societies
電機之夜熱音樂團的執行長 英文營的學術股長 AISEC海外實習計劃 YEF國際青年創業領袖計畫
Build Your Professional Network
Click icon on the company page or under talent search engine to start the conversation.