Welcome !!!

Hi, my name is Choo Jai Wei

  • An AI expertise who has more than 4.5-years working experience in applying different state-of-the-art AI solutions in the commercialized products including Manufacturing defect detection software, Mobile phone camera application, AI on edge devices etc.
  • I'm interested in tech especially in the field of AI and Data Science, I take pride in my ability to leverage different technologies to cater to the unique needs of each projects.
  • I'm excited about the opportunity to contribute my diverse AI skill set to your esteemed organization and to continue pushing the boundaries of innovation in the field of AI.
  • In my free time, I enjoy swimming and challenging myself in AI/ML competitions.

Skills

  • 👨‍💻Programming :
    | Python | C / C++ | SQL/noSQL |
  • 📚Framework/Library :
    | Pytorch | Tensorflow | Scikit-learn | XGBoost | LightGBM | OpenCV | NLTK | Hadoop | Postgres | MongoDB | Elasticsearch | LangChain | LangGraph |
  • 💻Devops Tools :
    | Docker | Git | Linux | GCP | Azure | Redis | Android |

Selected Commercialized Projects

Some of my accomplisment and projects I have worked on. Click the project title to understand more.

V2X Edge Computing Solution

Implemented different pruned and quantitized CNN models for real-time 3D object detection and segmentation with accuracy over 90%. Focused on Vehicle Identification and Traffic Obstacles Segmentation.

Skill stacks: Python, C++, Docker, ONNX, 2D/3D representation, Multi-task learning, Semi-supervised learning

Mobile Device's Camera Application Solution

Implemented different lightweighted CNN-based and transformer-based models for real-time scene recognition and 2D/3D humanpose estimation. Focused on Scene Recognition/Classification and AI Single Take (record & capture the best moment).

Skill stacks: Python, C++, tflite, 2D/3D representation, Image processing, Android Studio (adb)

Smart Manufacturing Solution - AI Inspection System

Developed and Implemented various state-of-the-art AI models to automate manufactuing inspection processes with auto buyoff defected products Focused on Image Similarity/Matching model, OCR and API development.

Skill stacks: Python, Restful API, Multiprocessing/Multithreaded, Postgres, ETL, Redis, Docker, Redis

Selected Self / University Projects

Microsoft Malware Prediction

A data science competition that hosted by Microsoft to predict if a machine will soon be hit with malware. Our solution landed 25th out of 2k teams on Kaggle. A simple solution summary by our team leader here.

Skill stacks: Python, Data Mining, Feature Engineering, Predictive Modelling

Maxis's Consultancy Project

A consultancy project that completed during my Master's program. I have helped the company to reduce operation cost in 5G infrastructure by building a predictive analytic dashboard that operate as a web-application and hosted on GCP virtual machine. The dashboard is taking data real-time data from GCP's BigQuery via a ETL pipeline and perform power consumption forecast based on the historical consumption.

Skill stacks: Python, Dash/Plotly, MongoDB, Spark, GCP, Time-series forecasting, ARIMA/RNN/LSTM

MobileViT Pytorch Minimal Implementation

Minimally reproduce mobile-based Transformer model (v1, v2 and v3) based on Apple's and Micron's research paper. More detail could visit my Github repo.

Skill stacks: Python, Transformer architecture, MobileNet, Vision Transformer