ResumeCreate your job-winning resume using our free resume builder.
PortfolioShowcase your skills and projects with a professional portfolio.
Cake AIUtilize our Cake AI-powered products to accelerate job search and talent match.ResumeCreate your job-winning resume using our free resume builder.Resume BuilderMake a resume for free.Resume TemplatesAccess our extensive library of professional & ready-to-use templates.Resume ExamplesGet inspired by real resume examples to create your own.Occupation GuideAccess resume writing guides tailored for different professions.Resume HelpGet expert advice on all things resume from our team of recruitment specialists.Cake AI Resume CheckerNewAnalyzes your resume using Cake AI for detailed feedback.- Success StoriesBusiness ExcellenceAbout CakeFeatured Reads
Cake Talent Search
The Most Lightweight and Effective Recruiting Plan
Search resumes and take the initiative to contact job applicants for higher recruiting efficiency. The Choice of Hundreds of Companies.
- Browse all search results
- Unlimited access to start new conversations
- Resumes accessible for only paid companies
- View users’ email address & phone numbers
賴泳瑄(Aimons)
Software Engineer
0986029578
大家好，我在大學期間創業開設飲料店持續經營5年，曾經同時管理2間店與12位工讀生，後來因為對於資訊的興趣繼續讀碩士，在碩士期間研究並學習AI與各種語言工具配合過5個科技部計畫，畢業後在東海擔任兼任講師與計畫助理，短暫一年間有在皚霖資訊開發醫療軟體系統，後續與廠商配合開發AI軟體並加入至黑瘋科技，至今在黑瘋科技期間開發過IOS/ANDROID APP、WEB、2D/3D GAME、LINUX設備都有配合過AI做出創新應用。
- AI軟體開發
- Web開發並上架
- RESTful Api 串接
- Unity遊戲開發並上架
- Linux設備開發
- Android, IOS App 軟體開發並上架
技能
前端
- Andorid Kotlin
- IOS Swift
- WEB Vue, Html, PHP, Javascript, Css, Ajax
- Linux Tkinter
- Unity
- UI/UX Photoshop, illustrator, Blender, SketchUp
- CI/CD Git
後端
- Language Python, Php, Javascript
- Host Services AWS, GCP
- Framework Laravel , Flask, Django
- Database MySQL, MangoDB, Firebase
- Container Docker
工作經歷
技術部主管 • 黑瘋科技股份有限公司
十二月 2019 - Present
研發beacon整體系統與產品、三軸整體系統與產品、九軸整體系統與產品，多種人工智慧系列產品。
管理人數 : 4, Android:1人, IOS:1人, Backend:1人, 設備: 1人
AI演算法撰寫與銜接
Android / IOS 藍牙底層SDK銜接
Android / IOS 專案框架及工程師疑難排解
APP WEB UX 規劃
第三方元件整合
Google Play / Apple Store 上架
兼任講師 • 東海大學
一月 2018 - Present
主要教學C、C++、JAVA、PYTHON、物聯網、資料科學、機器學習、深度學習等等。
科技部計畫研究助理 • 東海大學
一月 2018 - 十二月 2019
協助撰寫科技部計畫
與林口長庚醫院羅慧夫團隊配合製作AI齒顎矯正預測。
與台中榮總疾分師團隊配合製作AI疾分師醫囑分類ICD10。
與台中榮總配合製作X光片圖像辨識拍攝正確性。
技術部副理 • 皚霖資訊有限公司
一月 2018 - 一月 2019
協助醫療系統開發與測試
LINUX設備架構與開發
Web 開發上架維護
API串接
創業 • 歐趴飲料店
九月 2012 - 一月 2018
大學畢業後，開始自行創業，對於創業點子很多的我，希望能做出與別人有所不同並且試著實行，畢業就開始在東海商圈找尋可以承租的地點，開創改良飲料跟配料，並且以行動攤車將經營手搖茶，經過半年的努力生意開始好轉，經生意步入穩定後與工讀生，接著開始學習如何管理跟調配人力，工讀生的班次調配等等都花費相當大的心力，尤其是突然的請假的調班，還有各個員工不同習性的工作教學跟處理事情的態度教學，教持續經營了五年之久，後來因同時專注於碩士論文實作，時間無法調配，因而停止此項經營。
作品集
IMS員工管理系統
概要:
使用beacon技術綁定員工訊息，並使用樹梅派蒐集資訊，可以在非接觸環境下紀錄員工出勤、區域史、區域接觸史，並配合紅外線體溫計與環境感測器，可及時知道員工體溫狀況與室內外環境數據。
 |  |  |
使用技術:
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
ARES運動體感鞋
概要:
九軸感測器封裝至運動鞋內，配合運動軌跡辨識出使用者的左右腳步數、移動距離、運動軌跡、著地時間、騰空時間、內外旋傾角、著地姿勢、雙腳力道、有效力道、無效力道、配速、步頻等等運動資訊，藉由AI作數據分析給出評分與運動建議。正常測試使用體感芯片每1個月充電一次即可。
 |  |  |  |  |
使用技術:
bluetooth andorid/ios SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
ILOLLY幼稚園管理系統
概要:
使用beacon技術綁定幼稚園孩童，並使用樹莓派蒐集資訊，可以在非接觸環境下紀錄孩童出勤、區域史、區域接觸史，並配合紅外線體溫計與環境感測器，可及時知道孩童體溫狀況與室內外環境數據，還設置家長與孩童的溝通看板。
 |  |
使用技術:
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
TRAQ記步鞋
概要:
三軸感測器封裝至鞋內，不須充電可使用3-5年，配合記步演算法並傳輸資訊至APP，使用者無需24小時攜帶手機在身邊，鞋子本身的芯片會記錄7天資訊，只要7天內有靠近手機就會回傳7天資訊至APP並回傳雲端後台。
 |  |  |
使用技術:
bluetooth andorid/ios SDK、kotlin、swift、firebase、aws cloud、php laravel、python、SQLite、mySQL
ISPORT兒童體感遊戲平台
概要:
使用九軸感測器製作多種體感遊戲多達50種，平台內設置帳號可記錄各個體感遊戲遊玩之紀錄，並依據整體紀錄分析上肢核心下肢各部位的運動量，細部分析包含肌力、耐力、柔軟度、爆發力、平衡、敏捷、協調等，製作出目標達成指標並與同學齡孩童排行，如有不足的部分平台會做推薦，讓孩童可以輕鬆玩遊戲並享受健康的樂趣。
 |  |
使用技術:
bluetooth andorid/ios/windows SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
智慧體感魔法棒
概要:
九軸感測器封裝至魔法棒內，訓練空間中軌跡路徑，在空中畫出單一字母或數字，依據畫出的內容呈現出對應的效果，目前使用魔法遊戲當作DEMO。
使用技術:
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
面料辨識linebot
概要:
建置並拍攝面料工廠內200張資料，接著訓練出面料特徵，使用者可以使用圖片或拍攝傳入linebot，經過API傳輸至後台辨識並回傳資料庫內所有面料之相似度，找出相同或相似高的面料，使業務、客戶、商品開發人員可以快速了解可使用的庫存位置與數量或各個可代替的面料位置與數量，也可以直接使用文字描述做面料快速搜尋 。
 |  |  |  |  |
使用技術:
linebot、webhook、dialogflow、python、SQL、keras、tensorflow、flask、opencv、ngrok
設備管理系統
概要:
使用beacon技術綁定人員與設備訊息，並使用樹梅派蒐集資訊 ，可以設置設備歸屬區域與管理負責人，APP可即時顯示所有設備存在位置與狀態，設備自動會被紀錄區域史軌跡與借出歸還紀錄，可以免去大量人力管理與負擔，系統內也可以設置設備保養紀錄與需保養警示。
 |  |  |
 |  |  |
使用技術:
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、MySQL
智慧體感羽球拍
概要:
九軸感測器封裝至羽球拍內，可紀錄每場羽球的所有內容，包含揮拍次數、回擊方式(正拍反拍殺球吊球)、最大速度、平均速度、揮拍效率、揮拍軌跡等，可依據所有內容給予揮拍建議與訓練方式。
 |  |  |
使用技術:
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
ATS樣品鞋管理系統
概要:
使用beacon技術綁定樣品鞋，並使用樹莓派蒐集資訊與掃描平台，綁定過的樣品鞋可以在ATS掃描平台上顯示出該鞋子所有資訊，每當樣品鞋在特定情況移出樣品室後台網頁會自動做紀錄並發出緊示。
使用技術:
ibeacon、ble linux SDK、Linux、firebase 、Vue、aws cloud、php laravel、MySQL
AI疾分師醫囑分類ICD10
概要:
在醫院體系內都會有一群疾分師專門藉由醫生的診斷與醫囑做疾病分類，往往訓練疾分師示非常耗時而且稀少的，除了需要學習大量的專業醫療知識也需時常更新每年度的疾病分類，我們與台中榮總簽署病患資料保密協議，使用10年間並符合ICD10的病患資料訓練，訓練Bio BERT神經網路並更改transformer機制做出前5大疾病分類編碼的可能性，方便疾分師做快速篩選免除大量查找的時間，準確度高達7成6。
 |  |
技術:
python、ngrok、flask、MySQL、keras、tensorflow、Django、Bio BERT
AI齒顎矯正預測
概要:
齒顎矯正是一個包含外科手術、牙醫、美容外科三方面專業人員整合的技術，時常在矯正的過程中需要三方人員多次討論與檢討制定出齒顎矯正錨點(32個定位錨點)，一來一回的討論時間往往好幾個月計算，我們與林口長庚羅慧夫齒顎矯正團隊簽署資料保密協議，使用5年間矯正前與矯正後的3D錨點資料做訓練，訓練三組個別的神經網路，外科手術訓練鄰近點，牙醫訓練咬合契合度，美容外科訓練左右對稱性與黃金比例，最後embedding model神經網路整合出三方面知識的最後預測錨點，方便齒顎矯正討論縮短至1至2個禮拜，預測錨點的結果與醫生最後決定的錨點重疊率高達8成。
 |
技術:
python、ngrok、flask、MySQL、keras、tensorflow、Django、embedding model
X光片圖像辨識拍攝正確性
概要:
X光拍攝依據醫生醫囑決定拍攝部位，一間醫院不會設置過多X光機，因為操作人員的疏失或大量病患排隊X光機導致拍攝錯誤發生，往往病患需要來回診間多次才會知道拍攝錯誤，如果能在第一時間依據醫生醫囑與拍攝照片部位做比對，可以大幅降低拍攝錯誤情況發生，減少醫院X光機塞車的情況與醫院醫師之負擔，我們與台中榮總簽署病患X光照片保密協議，一開始在X光機上架設照相機蒐集300多張X光拍攝前照片與X光照，使用yolo神經網路做特徵提取並分析，可即時在拍攝X光照前辨識出部位，預測的部位與真實拍攝X光的部位準確率高達9成4。
 |  |
技術:
python、ngrok、flask、MySQL、keras、tensorflow、Django、Yolo
學歷
2015 - 2018
Tunghai University
資訊工程所
賴泳瑄(Aimons Lai)
Software Engineer
0986029578
Hello everyone, I started a Drink store and ran it for 5 years during college. I managed 2 stores and 12 parttime students at the same time. Later, I continued to study for a master's degree because of my interest in information. During the master's period, I researched and learned AI and different programing language tools. Cooperated with 5 Ministry of Science and Technology projects. After graduation, I worked as a part-time lecturer and planning assistant in 東海 University. For a short period of one year, I developed medical software systems at 皚霖 Technology Company. Later, I cooperated with manufacturers to develop AI software and joined AIHAVEFUN Technology Company. Now I developed IOS/ANDROID APP, WEB, 2D/3D GAME, and LINUX devices, all of which have cooperated with AI to make innovative applications.
- AI software development
- Web development and Publishment
- RESTful Api development
- Unity development and Publishment
- Linux device development
- Android, IOS App software development and Publishment
SKILL
Front-End
- Andorid Kotlin
- IOS Swift
- WEB Vue, Html, PHP, Javascript, Css, Ajax
- Linux Tkinter
- Unity
- UI/UX Photoshop, illustrator, Blender, SketchUp
- CI/CD Git
Back-End
- Language Python, Php, Javascript
- Host Services AWS, GCP
- Framework Laravel , Flask, Django
- Database MySQL, MangoDB, Firebase
- Container Docker
EXPERIENCE
IT Supervisor • AIHAVEFUN Technology Co., Ltd.
December 2019 - Present
Research and develop beacon overall systems and products, three-axis overall systems and products, nine-axis overall systems and products, and a variety of artificial intelligence series products.
Number of managers: 4, Android: 1, IOS: 1, Backend: 1, Device: 1
AI algorithm development
Android / IOS Bluetooth SDK integration
Android/IOS project framework and engineer troubleshooting
APP WEB UX planning
Third-party component integration
Publish on Google Play / Apple Store
Part-time Lecturer • Tunghai University
January 2018 - Present
Mainly teaching C, C++, JAVA, PYTHON, Internet of Things, data science, machine learning, deep learning, etc.
Research Assistant • Tunghai University
January 2018 - December 2019
Assist in writing plans for the Ministry of Science and Technology
Collaborated with Luo Huifu’s team at Linkou Chang Gung Memorial Hospital to produce AI tooth and jaw correction predictions.
Worked with Taichung Veterans General Hospital’s disease diagnostic team to create AI disease diagnostician’s medical order classification ICD10.
Cooperate with Taichung Veterans General Hospital to produce X-ray images to identify the accuracy of shooting.
Assistant Manager of Technology Department • S.I-LINK CO., LTD.
January 2018 - January 2019
Assist with medical system development and testing
LINUX device architecture and development
Web development and maintenance
API connection
Startup • AllPass Drink Store
September 2012 - January 2018
After graduating from college, I started my own business. I had a lot of entrepreneurial ideas, so I hoped to be different from others and try to implement them. After graduation, I started looking for a place to rent in the Tunghai business district, and created improved beverages and ingredients. The mobile stall will operate hand-cranked tea. After half a year of hard work, the business started to improve. After the business stabilized, I started to learn how to manage and allocate manpower with the part-time students. It took a lot of effort to arrange the shifts of the part-time students. Especially the sudden shift, as well as the teaching of work habits and attitude of handling things with different employees. Continued for five years. Later, because it was focused on the implementation of the master's thesis at the same time, the time could not be allocated, so i was stopped this business.
Portfolio
IMS employee management system
Summary :
Use beacon technology to bind employee information and use raspberry pi to collect information. It can record employee attendance, regional history, and regional contact history in a non-contact environment. With infrared thermometers and environmental sensors, employees can know their temperature status and Indoor/outdoor environmental data in real time.
| | |
Technique :
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
ARES sports shoes
Summary :
The nine-axis sensor is packaged into the sneaker and can identify the user's left and right steps, movement distance, movement trajectory, landing time, flight time, internal and external rotation angle, landing posture, foot strength, effective strength, ineffectiveness strength, pace, cadence, etc. are analyzed using AI to provide scores and exercise suggestions. For normal testing, the somatosensory chip can be charged once per month.
| | | | |
Technique :
bluetooth andorid/ios SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
ILOLLY Kindergarten Management System
Summary :
Using beacon technology to bind kindergarten children and using Raspberry Pi to collect information, children's attendance, regional history, and regional contact history can be recorded in a non-contact environment. With infrared thermometers and environmental sensors, children's body temperature status and Indoor and outdoor environmental data, and communication boards for parents and children.
| |
Technique :
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
TRAQ step tracking shoes
Summary :
The three-axis sensor is packaged into the shoe and can be used for 3-5 years without charging. It cooperates with the step counting algorithm and transmits information to the APP. Users do not need to carry their mobile phones anytime. The chip in the shoe itself will record 7 days of information. , as long as someone is close to the phone within 7 days, 7 days of information will be sent back to the APP and sent back to the cloud background.
| | |
Technique :
bluetooth andorid/ios SDK、kotlin、swift、firebase、aws cloud、php laravel、python、SQLite、mySQL
ISPORT children's somatosensory game platform
Summary :
Use nine-axis sensors to create up to 50 types of various motion-sensing games. Set up an account on the platform to record the records of each motion-sensing game. Based on the overall record, the amount of exercise in each part of the upper limbs, core and lower limbs is analyzed. Detailed analysis includes muscle strength, endurance, Flexibility, explosive power, balance, agility, coordination, etc., create goal achievement indicators and rank them with children of the same age. If there are deficiencies, platforms will make recommendations, so that children can easily play games and enjoy healthy fun.
| |
Technique :
bluetooth andorid/ios/windows SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
Smart motion magic wand
Summary :
The nine-axis sensor is packaged in a magic wand, and training trajectory path. Draw a single letter or number in the air. The corresponding effect is displayed based on the drawn content. Ues game for DEMO.
Technique :
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
Leather identification linebot
Summary :
Create and photograph 200 pieces of data in the leather factory, and then train the leather characteristics. Users can use pictures or photos to input to the linebot, and transmit it to the backend through the API to identify and return the similarity of all leathers in the database to find the same or leather with high similarity. Allow business, customers, and product developers to quickly understand the available inventory location and quantity or the location and quantity of each replaceable leathers . They can also directly use text descriptions to quickly search for leathers .
| | | | |
Technique :
linebot、webhook、dialogflow、python、SQL、keras、tensorflow、flask、opencv、ngrok
Equipment management system
Summary :
Use beacon technology to bind personnel and equipment information, and use Raspberry Pi to collect information. You can set the equipment ownership area and management person. The APP can instantly display the location and status of all equipment, and the equipment will automatically be recorded in the regional history and loan-Return records.it can eliminate a lot of manpower management and burden. Equipment maintenance records and maintenance-required warnings can also be set in the system.
| | |
| | |
Technique :
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、MySQL
Smart motion badminton racket
Summary :
The nine-axis sensor is packaged into the badminton racket and can record all the content of each badminton game, including the number of swings, return methods (forehand and backhand smashes and lobs), maximum speed, average speed, swing efficiency, swing Trajectories, etc., and swing suggestions and training methods can be given based on all content.
| | |
Technique :
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
ATS sample shoe management system
Summary :
Use beacon technology to bind sample shoes, and use Raspberry Pi to collect information and scan platform. The bounded sample shoes can display all the information about the shoes on the ATS scanning platform. When the sample shoes are moved out from the sample room, the backend web page will automatically record it and issue an alert.
Technique :
ibeacon、ble linux SDK、Linux、firebase 、Vue、aws cloud、php laravel、MySQL
AI Disease Classifier ICD10
Summary :
In the hospital system, there will be a group of pathologists who specialize in classifying diseases based on doctors’ diagnoses and orders. Training of pathologists is often very time-consuming and rare. In addition to learning a lot of professional medical knowledge, it also needs to be updated every year. For disease classification, we signed a patient data confidentiality agreement with Taichung Veterans General Hospital. We used patient data from 10 years and complied with ICD10 to train the Bio BERT neural network and changed the transformer mechanism to make it possible to code the top five disease classifications, which is convenient. The disease analyzer can perform quick screening without a lot of search time, and the accuracy is 76%.
| |
Technique :
python、ngrok、flask、MySQL、keras、tensorflow、Django、Bio BERT
AI Orthodontic correction prediction
Summary :
Orthodontic correction is a technology that involves the integration of professionals from surgery, dentistry, and aesthetic medicine. During the correction process, it often requires multiple discussions and reviews by the three parties to formulate orthodontic correction anchor points (32 positioning anchor points). A lot of discussions often take several months. We signed a data confidentiality agreement with the Linkou Chang Geng Luo Huifu Orthodontic Correction Team, and used 3D anchor point data before and after correction for five years to train three groups of individual neural networks. Path, surgery trains adjacent points, dentists train occlusion fit, aesthetic medicine trains left and right symmetry and the golden ratio, and finally the embedding model neural network integrates the final prediction anchor point of the three aspects of knowledge, which facilitates shortening the discussion of Orthodontic correction to 1 to 2 weeks, the overlap rate between the predicted anchor point and the anchor point finally decided by the doctor was 80%.
|
Technique :
python、ngrok、flask、MySQL、keras、tensorflow、Django、embedding model
Correctness of X-ray image recognition and photograph
Summary :
X-ray photograph is determined based on the doctor's orders. A hospital will not build too many X-ray machines. Photograph errors occur due to operator errors or a lot of patients queuing. Patients often need to go back multiple times. then will be aware of photograph errors. If it can compare the photograph and the doctor's order at the same time, it can reduce the occurrence of photograph errors, then reduce the traffic jam in the hospital and Waste of manpower. We signed a patient X-ray photo NDA with Taichung Veterans General Hospital . Initially, a camera was set up on the X-ray machine to collect more than 300 pre-X-ray photos and X-rays. The yolo neural network was used for feature extraction and analysis to identify the parts immediately after taking the X-rays photograph . the accuracy of the predicted parts and the actual X-ray parts is 94%.
| |
Technique :
python、ngrok、flask、MySQL、keras、tensorflow、Django、Yolo
Education
2015 - 2018
Tunghai University
Computer Science and Information Engineering
賴泳瑄(Aimons)
Software Engineer
0986029578
大家好，我在大學期間創業開設飲料店持續經營5年，曾經同時管理2間店與12位工讀生，後來因為對於資訊的興趣繼續讀碩士，在碩士期間研究並學習AI與各種語言工具配合過5個科技部計畫，畢業後在東海擔任兼任講師與計畫助理，短暫一年間有在皚霖資訊開發醫療軟體系統，後續與廠商配合開發AI軟體並加入至黑瘋科技，至今在黑瘋科技期間開發過IOS/ANDROID APP、WEB、2D/3D GAME、LINUX設備都有配合過AI做出創新應用。
- AI軟體開發
- Web開發並上架
- RESTful Api 串接
- Unity遊戲開發並上架
- Linux設備開發
- Android, IOS App 軟體開發並上架
技能
前端
- Andorid Kotlin
- IOS Swift
- WEB Vue, Html, PHP, Javascript, Css, Ajax
- Linux Tkinter
- Unity
- UI/UX Photoshop, illustrator, Blender, SketchUp
- CI/CD Git
後端
- Language Python, Php, Javascript
- Host Services AWS, GCP
- Framework Laravel , Flask, Django
- Database MySQL, MangoDB, Firebase
- Container Docker
工作經歷
技術部主管 • 黑瘋科技股份有限公司
十二月 2019 - Present
研發beacon整體系統與產品、三軸整體系統與產品、九軸整體系統與產品，多種人工智慧系列產品。
管理人數 : 4, Android:1人, IOS:1人, Backend:1人, 設備: 1人
AI演算法撰寫與銜接
Android / IOS 藍牙底層SDK銜接
Android / IOS 專案框架及工程師疑難排解
APP WEB UX 規劃
第三方元件整合
Google Play / Apple Store 上架
兼任講師 • 東海大學
一月 2018 - Present
主要教學C、C++、JAVA、PYTHON、物聯網、資料科學、機器學習、深度學習等等。
科技部計畫研究助理 • 東海大學
一月 2018 - 十二月 2019
協助撰寫科技部計畫
與林口長庚醫院羅慧夫團隊配合製作AI齒顎矯正預測。
與台中榮總疾分師團隊配合製作AI疾分師醫囑分類ICD10。
與台中榮總配合製作X光片圖像辨識拍攝正確性。
技術部副理 • 皚霖資訊有限公司
一月 2018 - 一月 2019
協助醫療系統開發與測試
LINUX設備架構與開發
Web 開發上架維護
API串接
創業 • 歐趴飲料店
九月 2012 - 一月 2018
大學畢業後，開始自行創業，對於創業點子很多的我，希望能做出與別人有所不同並且試著實行，畢業就開始在東海商圈找尋可以承租的地點，開創改良飲料跟配料，並且以行動攤車將經營手搖茶，經過半年的努力生意開始好轉，經生意步入穩定後與工讀生，接著開始學習如何管理跟調配人力，工讀生的班次調配等等都花費相當大的心力，尤其是突然的請假的調班，還有各個員工不同習性的工作教學跟處理事情的態度教學，教持續經營了五年之久，後來因同時專注於碩士論文實作，時間無法調配，因而停止此項經營。
作品集
IMS員工管理系統
概要:
使用beacon技術綁定員工訊息，並使用樹梅派蒐集資訊，可以在非接觸環境下紀錄員工出勤、區域史、區域接觸史，並配合紅外線體溫計與環境感測器，可及時知道員工體溫狀況與室內外環境數據。
| | |
使用技術:
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
ARES運動體感鞋
概要:
九軸感測器封裝至運動鞋內，配合運動軌跡辨識出使用者的左右腳步數、移動距離、運動軌跡、著地時間、騰空時間、內外旋傾角、著地姿勢、雙腳力道、有效力道、無效力道、配速、步頻等等運動資訊，藉由AI作數據分析給出評分與運動建議。正常測試使用體感芯片每1個月充電一次即可。
| | | | |
使用技術:
bluetooth andorid/ios SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
ILOLLY幼稚園管理系統
概要:
使用beacon技術綁定幼稚園孩童，並使用樹莓派蒐集資訊，可以在非接觸環境下紀錄孩童出勤、區域史、區域接觸史，並配合紅外線體溫計與環境感測器，可及時知道孩童體溫狀況與室內外環境數據，還設置家長與孩童的溝通看板。
| |
使用技術:
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
TRAQ記步鞋
概要:
三軸感測器封裝至鞋內，不須充電可使用3-5年，配合記步演算法並傳輸資訊至APP，使用者無需24小時攜帶手機在身邊，鞋子本身的芯片會記錄7天資訊，只要7天內有靠近手機就會回傳7天資訊至APP並回傳雲端後台。
| | |
使用技術:
bluetooth andorid/ios SDK、kotlin、swift、firebase、aws cloud、php laravel、python、SQLite、mySQL
ISPORT兒童體感遊戲平台
概要:
使用九軸感測器製作多種體感遊戲多達50種，平台內設置帳號可記錄各個體感遊戲遊玩之紀錄，並依據整體紀錄分析上肢核心下肢各部位的運動量，細部分析包含肌力、耐力、柔軟度、爆發力、平衡、敏捷、協調等，製作出目標達成指標並與同學齡孩童排行，如有不足的部分平台會做推薦，讓孩童可以輕鬆玩遊戲並享受健康的樂趣。
| |
使用技術:
bluetooth andorid/ios/windows SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
智慧體感魔法棒
概要:
九軸感測器封裝至魔法棒內，訓練空間中軌跡路徑，在空中畫出單一字母或數字，依據畫出的內容呈現出對應的效果，目前使用魔法遊戲當作DEMO。
使用技術:
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
面料辨識linebot
概要:
建置並拍攝面料工廠內200張資料，接著訓練出面料特徵，使用者可以使用圖片或拍攝傳入linebot，經過API傳輸至後台辨識並回傳資料庫內所有面料之相似度，找出相同或相似高的面料，使業務、客戶、商品開發人員可以快速了解可使用的庫存位置與數量或各個可代替的面料位置與數量，也可以直接使用文字描述做面料快速搜尋 。
| | | | |
使用技術:
linebot、webhook、dialogflow、python、SQL、keras、tensorflow、flask、opencv、ngrok
設備管理系統
概要:
使用beacon技術綁定人員與設備訊息，並使用樹梅派蒐集資訊 ，可以設置設備歸屬區域與管理負責人，APP可即時顯示所有設備存在位置與狀態，設備自動會被紀錄區域史軌跡與借出歸還紀錄，可以免去大量人力管理與負擔，系統內也可以設置設備保養紀錄與需保養警示。
| | |
| | |
使用技術:
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、MySQL
智慧體感羽球拍
概要:
九軸感測器封裝至羽球拍內，可紀錄每場羽球的所有內容，包含揮拍次數、回擊方式(正拍反拍殺球吊球)、最大速度、平均速度、揮拍效率、揮拍軌跡等，可依據所有內容給予揮拍建議與訓練方式。
| | |
使用技術:
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
ATS樣品鞋管理系統
概要:
使用beacon技術綁定樣品鞋，並使用樹莓派蒐集資訊與掃描平台，綁定過的樣品鞋可以在ATS掃描平台上顯示出該鞋子所有資訊，每當樣品鞋在特定情況移出樣品室後台網頁會自動做紀錄並發出緊示。
使用技術:
ibeacon、ble linux SDK、Linux、firebase 、Vue、aws cloud、php laravel、MySQL
AI疾分師醫囑分類ICD10
概要:
在醫院體系內都會有一群疾分師專門藉由醫生的診斷與醫囑做疾病分類，往往訓練疾分師示非常耗時而且稀少的，除了需要學習大量的專業醫療知識也需時常更新每年度的疾病分類，我們與台中榮總簽署病患資料保密協議，使用10年間並符合ICD10的病患資料訓練，訓練Bio BERT神經網路並更改transformer機制做出前5大疾病分類編碼的可能性，方便疾分師做快速篩選免除大量查找的時間，準確度高達7成6。
| |
技術:
python、ngrok、flask、MySQL、keras、tensorflow、Django、Bio BERT
AI齒顎矯正預測
概要:
齒顎矯正是一個包含外科手術、牙醫、美容外科三方面專業人員整合的技術，時常在矯正的過程中需要三方人員多次討論與檢討制定出齒顎矯正錨點(32個定位錨點)，一來一回的討論時間往往好幾個月計算，我們與林口長庚羅慧夫齒顎矯正團隊簽署資料保密協議，使用5年間矯正前與矯正後的3D錨點資料做訓練，訓練三組個別的神經網路，外科手術訓練鄰近點，牙醫訓練咬合契合度，美容外科訓練左右對稱性與黃金比例，最後embedding model神經網路整合出三方面知識的最後預測錨點，方便齒顎矯正討論縮短至1至2個禮拜，預測錨點的結果與醫生最後決定的錨點重疊率高達8成。
|
技術:
python、ngrok、flask、MySQL、keras、tensorflow、Django、embedding model
X光片圖像辨識拍攝正確性
概要:
X光拍攝依據醫生醫囑決定拍攝部位，一間醫院不會設置過多X光機，因為操作人員的疏失或大量病患排隊X光機導致拍攝錯誤發生，往往病患需要來回診間多次才會知道拍攝錯誤，如果能在第一時間依據醫生醫囑與拍攝照片部位做比對，可以大幅降低拍攝錯誤情況發生，減少醫院X光機塞車的情況與醫院醫師之負擔，我們與台中榮總簽署病患X光照片保密協議，一開始在X光機上架設照相機蒐集300多張X光拍攝前照片與X光照，使用yolo神經網路做特徵提取並分析，可即時在拍攝X光照前辨識出部位，預測的部位與真實拍攝X光的部位準確率高達9成4。
| |
技術:
python、ngrok、flask、MySQL、keras、tensorflow、Django、Yolo
學歷
2015 - 2018
Tunghai University
資訊工程所
賴泳瑄(Aimons Lai)
Software Engineer
0986029578
Hello everyone, I started a Drink store and ran it for 5 years during college. I managed 2 stores and 12 parttime students at the same time. Later, I continued to study for a master's degree because of my interest in information. During the master's period, I researched and learned AI and different programing language tools. Cooperated with 5 Ministry of Science and Technology projects. After graduation, I worked as a part-time lecturer and planning assistant in 東海 University. For a short period of one year, I developed medical software systems at 皚霖 Technology Company. Later, I cooperated with manufacturers to develop AI software and joined AIHAVEFUN Technology Company. Now I developed IOS/ANDROID APP, WEB, 2D/3D GAME, and LINUX devices, all of which have cooperated with AI to make innovative applications.
- AI software development
- Web development and Publishment
- RESTful Api development
- Unity development and Publishment
- Linux device development
- Android, IOS App software development and Publishment
SKILL
Front-End
- Andorid Kotlin
- IOS Swift
- WEB Vue, Html, PHP, Javascript, Css, Ajax
- Linux Tkinter
- Unity
- UI/UX Photoshop, illustrator, Blender, SketchUp
- CI/CD Git
Back-End
- Language Python, Php, Javascript
- Host Services AWS, GCP
- Framework Laravel , Flask, Django
- Database MySQL, MangoDB, Firebase
- Container Docker
EXPERIENCE
IT Supervisor • AIHAVEFUN Technology Co., Ltd.
December 2019 - Present
Research and develop beacon overall systems and products, three-axis overall systems and products, nine-axis overall systems and products, and a variety of artificial intelligence series products.
Number of managers: 4, Android: 1, IOS: 1, Backend: 1, Device: 1
AI algorithm development
Android / IOS Bluetooth SDK integration
Android/IOS project framework and engineer troubleshooting
APP WEB UX planning
Third-party component integration
Publish on Google Play / Apple Store
Part-time Lecturer • Tunghai University
January 2018 - Present
Mainly teaching C, C++, JAVA, PYTHON, Internet of Things, data science, machine learning, deep learning, etc.
Research Assistant • Tunghai University
January 2018 - December 2019
Assist in writing plans for the Ministry of Science and Technology
Collaborated with Luo Huifu’s team at Linkou Chang Gung Memorial Hospital to produce AI tooth and jaw correction predictions.
Worked with Taichung Veterans General Hospital’s disease diagnostic team to create AI disease diagnostician’s medical order classification ICD10.
Cooperate with Taichung Veterans General Hospital to produce X-ray images to identify the accuracy of shooting.
Assistant Manager of Technology Department • S.I-LINK CO., LTD.
January 2018 - January 2019
Assist with medical system development and testing
LINUX device architecture and development
Web development and maintenance
API connection
Startup • AllPass Drink Store
September 2012 - January 2018
After graduating from college, I started my own business. I had a lot of entrepreneurial ideas, so I hoped to be different from others and try to implement them. After graduation, I started looking for a place to rent in the Tunghai business district, and created improved beverages and ingredients. The mobile stall will operate hand-cranked tea. After half a year of hard work, the business started to improve. After the business stabilized, I started to learn how to manage and allocate manpower with the part-time students. It took a lot of effort to arrange the shifts of the part-time students. Especially the sudden shift, as well as the teaching of work habits and attitude of handling things with different employees. Continued for five years. Later, because it was focused on the implementation of the master's thesis at the same time, the time could not be allocated, so i was stopped this business.
Portfolio
IMS employee management system
Summary :
Use beacon technology to bind employee information and use raspberry pi to collect information. It can record employee attendance, regional history, and regional contact history in a non-contact environment. With infrared thermometers and environmental sensors, employees can know their temperature status and Indoor/outdoor environmental data in real time.
| | |
Technique :
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
ARES sports shoes
Summary :
The nine-axis sensor is packaged into the sneaker and can identify the user's left and right steps, movement distance, movement trajectory, landing time, flight time, internal and external rotation angle, landing posture, foot strength, effective strength, ineffectiveness strength, pace, cadence, etc. are analyzed using AI to provide scores and exercise suggestions. For normal testing, the somatosensory chip can be charged once per month.
| | | | |
Technique :
bluetooth andorid/ios SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
ILOLLY Kindergarten Management System
Summary :
Using beacon technology to bind kindergarten children and using Raspberry Pi to collect information, children's attendance, regional history, and regional contact history can be recorded in a non-contact environment. With infrared thermometers and environmental sensors, children's body temperature status and Indoor and outdoor environmental data, and communication boards for parents and children.
| |
Technique :
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、mySQL
TRAQ step tracking shoes
Summary :
The three-axis sensor is packaged into the shoe and can be used for 3-5 years without charging. It cooperates with the step counting algorithm and transmits information to the APP. Users do not need to carry their mobile phones anytime. The chip in the shoe itself will record 7 days of information. , as long as someone is close to the phone within 7 days, 7 days of information will be sent back to the APP and sent back to the cloud background.
| | |
Technique :
bluetooth andorid/ios SDK、kotlin、swift、firebase、aws cloud、php laravel、python、SQLite、mySQL
ISPORT children's somatosensory game platform
Summary :
Use nine-axis sensors to create up to 50 types of various motion-sensing games. Set up an account on the platform to record the records of each motion-sensing game. Based on the overall record, the amount of exercise in each part of the upper limbs, core and lower limbs is analyzed. Detailed analysis includes muscle strength, endurance, Flexibility, explosive power, balance, agility, coordination, etc., create goal achievement indicators and rank them with children of the same age. If there are deficiencies, platforms will make recommendations, so that children can easily play games and enjoy healthy fun.
| |
Technique :
bluetooth andorid/ios/windows SDK、Unity、aws cloud、php laravel、python、keras、tensorflow、SQLite、mySQL
Smart motion magic wand
Summary :
The nine-axis sensor is packaged in a magic wand, and training trajectory path. Draw a single letter or number in the air. The corresponding effect is displayed based on the drawn content. Ues game for DEMO.
Technique :
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
Leather identification linebot
Summary :
Create and photograph 200 pieces of data in the leather factory, and then train the leather characteristics. Users can use pictures or photos to input to the linebot, and transmit it to the backend through the API to identify and return the similarity of all leathers in the database to find the same or leather with high similarity. Allow business, customers, and product developers to quickly understand the available inventory location and quantity or the location and quantity of each replaceable leathers . They can also directly use text descriptions to quickly search for leathers .
| | | | |
Technique :
linebot、webhook、dialogflow、python、SQL、keras、tensorflow、flask、opencv、ngrok
Equipment management system
Summary :
Use beacon technology to bind personnel and equipment information, and use Raspberry Pi to collect information. You can set the equipment ownership area and management person. The APP can instantly display the location and status of all equipment, and the equipment will automatically be recorded in the regional history and loan-Return records.it can eliminate a lot of manpower management and burden. Equipment maintenance records and maintenance-required warnings can also be set in the system.
| | |
| | |
Technique :
ibeacon、ble android/ios/linux SDK、kotlin、swift、firebase、Linux、python、aws cloud 、php laravel、MySQL
Smart motion badminton racket
Summary :
The nine-axis sensor is packaged into the badminton racket and can record all the content of each badminton game, including the number of swings, return methods (forehand and backhand smashes and lobs), maximum speed, average speed, swing efficiency, swing Trajectories, etc., and swing suggestions and training methods can be given based on all content.
| | |
Technique :
bluetooth andorid/ios/windows SDK、Unity 、python、keras、tensorflow、SQLite
ATS sample shoe management system
Summary :
Use beacon technology to bind sample shoes, and use Raspberry Pi to collect information and scan platform. The bounded sample shoes can display all the information about the shoes on the ATS scanning platform. When the sample shoes are moved out from the sample room, the backend web page will automatically record it and issue an alert.
Technique :
ibeacon、ble linux SDK、Linux、firebase 、Vue、aws cloud、php laravel、MySQL
AI Disease Classifier ICD10
Summary :
In the hospital system, there will be a group of pathologists who specialize in classifying diseases based on doctors’ diagnoses and orders. Training of pathologists is often very time-consuming and rare. In addition to learning a lot of professional medical knowledge, it also needs to be updated every year. For disease classification, we signed a patient data confidentiality agreement with Taichung Veterans General Hospital. We used patient data from 10 years and complied with ICD10 to train the Bio BERT neural network and changed the transformer mechanism to make it possible to code the top five disease classifications, which is convenient. The disease analyzer can perform quick screening without a lot of search time, and the accuracy is 76%.
| |
Technique :
python、ngrok、flask、MySQL、keras、tensorflow、Django、Bio BERT
AI Orthodontic correction prediction
Summary :
Orthodontic correction is a technology that involves the integration of professionals from surgery, dentistry, and aesthetic medicine. During the correction process, it often requires multiple discussions and reviews by the three parties to formulate orthodontic correction anchor points (32 positioning anchor points). A lot of discussions often take several months. We signed a data confidentiality agreement with the Linkou Chang Geng Luo Huifu Orthodontic Correction Team, and used 3D anchor point data before and after correction for five years to train three groups of individual neural networks. Path, surgery trains adjacent points, dentists train occlusion fit, aesthetic medicine trains left and right symmetry and the golden ratio, and finally the embedding model neural network integrates the final prediction anchor point of the three aspects of knowledge, which facilitates shortening the discussion of Orthodontic correction to 1 to 2 weeks, the overlap rate between the predicted anchor point and the anchor point finally decided by the doctor was 80%.
|
Technique :
python、ngrok、flask、MySQL、keras、tensorflow、Django、embedding model
Correctness of X-ray image recognition and photograph
Summary :
X-ray photograph is determined based on the doctor's orders. A hospital will not build too many X-ray machines. Photograph errors occur due to operator errors or a lot of patients queuing. Patients often need to go back multiple times. then will be aware of photograph errors. If it can compare the photograph and the doctor's order at the same time, it can reduce the occurrence of photograph errors, then reduce the traffic jam in the hospital and Waste of manpower. We signed a patient X-ray photo NDA with Taichung Veterans General Hospital . Initially, a camera was set up on the X-ray machine to collect more than 300 pre-X-ray photos and X-rays. The yolo neural network was used for feature extraction and analysis to identify the parts immediately after taking the X-rays photograph . the accuracy of the predicted parts and the actual X-ray parts is 94%.
| |
Technique :
python、ngrok、flask、MySQL、keras、tensorflow、Django、Yolo
Education
2015 - 2018