CHAI

YOUXIANG

Self Portrait
Hey! I'm a
Self Portrait
About me

My name is Chai Youxiang and I'm bachelor of Computer Science from Nanyang Technological University, Singapore. I'm currently living my dream of being an AI Fullstack Developer!.

My goal is to create software that is not only functional, but engaging and user-oriented. I strongly believe that there is a place for Machine Learning and Artificial Intelligence in creating experiences that resonate with the user and applications that are cleverly intuitive and easy to use.

Skills & Technologies

Front-End

Front-End

Javascript
Typescript
HTML
CSS
ReactJS
NextJS
Framer Motion
Figma
TailwindCSS

Back-End

Back-End

PostgreSQL
GraphQL
Python
NodeJS
Firebase

DevOps

DevOps

NGINX
Docker
Linux
Git
Github

Languages & Libraries

Languages & Libraries

C
Golang
Pytorch
SciKit Learn
OpenCV
Pandas
Numpy

My journey so far

FinTech Intern

Ernst & Young, Singapore

ASEAN FinTech Intern

August 2017 - Feburary 2018

  • Collaborated on and helped maintain a database of 1,000 Global Fintech Startups and 500 Investors for EY’s Investor Summit during the Singapore Fintech Festival 2017
  • Initiated and automated Startup and Investor matching using SQL

Software Engineering Intern

Ministry of Home Affairs, Singapore

Software Engineering Intern

May 2022 - December 2022

  • Designed and developed the front-end UI/UX for a mobile/web cross-platform application to track completion rates and progress of users within the department
  • Leverages tools such as Figma for wireframing and collaboration; Node.js, React.js, Typescript and the Ionic framework for front-end development
  • Employed a CI/CD pipeline using Microsoft Azure DevOps, GitHub and deployed the application onto Azure Kubernetes Service

Bachelor's Degree Graduation

Nanyang Technological University, Singapore

Bachelor of Engineering in Computer Science

August 2020 - January 2024

  • Elective focus - Artificial Intelligence, Data Science & Analytics

My projects

Deep Learning Methods with Less Supervision

Deep Learning Methods with Less Supervision

(Final Year Project)

Developed a framework for training segmentation models with weak forms of supervision, using foundational models such as Meta's Segment Anything Model for pseudo-label generation.

Fine-tuned a fully-connected-network with a ResNet50 backbone based on generated pseudo labels and incorporated Conditional Random Fields as a post-processing technique.

PythonPyTorchOpenCVSegment-Anything-Model
Developer Portfolio Website

Developer Portfolio Website

A personal web developer portfolio website created using React and NextJS, written in Typescript.

TSReactJSNext.JSFramer-Motion
Knowledge Distillation Exploration

Knowledge Distillation Exploration

Explored Knowledge Distillation using PyTorch models for classification on the Fashion-MNIST dataset

Experimented with distilling knowledge from ResNet as the teacher model to MobileNet as the student model using KL-Divergence as the loss function

PyTorchFashion-MNISTTorchvision

Let's get in touch!