Hi there, I'm Elston đź‘‹

I am a 20 year old AI Researcher, Entrepreneur and Full-Stack Developer from Singapore. I specialize in deep learning (CV and NLP) and am continuously exploring different fields of AI.

I mainly do research into deep learning such as reading newly published papers from conferences and looking into new fields of AI such as State Space Models, Graph Deep Learning and Reinforcement Learning as of right now.

I also do a lot of AI engineering work such as:
1. Building end-to-end pipelines and deploying them as well optimizing them. (e.g. porting models to C++ and TensorRT)
2. Using newly published frameworks and libraries to create cool projects
3. Participating in hackathons/competitions and applying state-of-the-art technology paired with full-stack and deployment methods

As for personal interests, I do Muay Thai as a hobby and I frequently work out. My academic interests also tend to vary and I've delved deep into quantum physics, scrum, devops, neuroscience and anything I find interesting tbh.

Avatar image
Recent Projects
View all Projects →
tensorrt

TensorRT Deployment For Defect Segmentation Model

Deep Learning Deployment C++

Wed Jan 01 2025

Created C++ SDK for deployment purposes of a defect segmentation model for A*STAR

efficientvod

Video object detection by location anticipation windows implementation

Deep Learning Video Object Detection

Thu Dec 05 2024

Created a windows support for the ICCV 2023 paper, 'Objects do not disappear: Video object detection by single-frame object location anticipation.'

gradient boost graph

Gradient Boosting Regressor From Scratch

Machine Learning

Tue Nov 19 2024

Creating gradient boosting regressor from scratch using stumps

transformer

Transformer Architecture From Scratch

NLP, Deep Learning

Tue Nov 12 2024

A small project to recreate the Transformer Architecture from scratch from the paper and train it on a language translation dataset

quantum circuit

Quantum Random Number Generator

Quantum Computing

Fri Sep 29 2023

A small project to simulate a random number generator but using quantum computations

Job Experiences

'Sup

4/1/2025 - 9/2/2025
'Sup logo.

Chief Technology Officer (Jun 2025 - Present)

Responsible for leading a multinational development team with members from the US, London, Italy and Singapore.

  • Scaled our platform and introduced an increase of ~83.7% CMAU to 90 in less than a week
  • Collaborated with external organizations, managed the team and logistics for the collaborations and introduced a 2x increase in userbase during one conference
  • In-charged of infrastructure and handled cyber attacks and data breaches

Founding Technical Lead (Apr 2025 - Jun 2025)

Responsible for leading a multinational development team of 5 from the US and Singapore to ensure flawless execution of pilot sessions across multiple schools in Singapore

  • Lead the development for a pilot sessions at CHIJ, SAJC and BuildingBlocs
  • Accumulated over 700 users on our live waitlist
  • Handled traffic and website for serving over 100 concurrent users on live production site
  • Implemented model/technology changes
  • Responsible for deployment of production site with continuous integration and development methods
  • Handled technology costs and maximized for cost-efficiency
  • Spearheaded collaboration efforts with clubs in Singapore Polytechnic and BuildingBlocs

Advanced Remanufacturing and Technology Centre, A*STAR

3/1/2024 - 7/4/2025
A*STAR logo.

Research Engineer (Jan 2025 - July 2025)

Full-time contracted AI Engineer for A*STAR ARTC Intelligent Product Verification (IPV) department reporting under the Deputy Group Manager.

  • Ported models to TensorRT engine and onnx for C++ runtime
  • Implemented model compression and sped models up by ~40% through methods like quantization etc
  • Developed and documented an SDK (Software Development Kit) in C++ for model deployment
  • Rapidly developed proof-of-concept solutions for managers and clients
  • Improved model performances by ~15% through optimization
  • Led the development and ensured the success of three client projects
  • Secured funding for projects
  • Led the development and closed multiple projects worth ~170k each with high customer satisfaction/feedback
  • Collected and annotated datasets from clients
  • Crafted comprehensive reports for group managers and clients
  • Undertook research, prototyping and finetuning of SOTA VLMs for real-world use-cases for multiple clients
  • Developed a full-stack data platform and containerised for usage by clients (converted frontend/backend code into machine code and containerised using podman)
  • Prototyped and researched Monocular Depth Estimation as well as light microscopy reconstruction with Physics Informed Diffusion Models

AI Research Engineer Intern (Mar 2024 - Jan 2025)

  • Doing research into state of the art models and architectures and reproducing papers for client projects as well as in-house research projects for ARTC IPV (Intelligent Product Verification).
  • Created end-to-end pipelines and models for clients and created proof-of-concept solutions for in-house research projects.
  • Created proof-of-concept defect detection solution that met KPIs for Hyundai Steel and managed to secure a contract and funding for the project.
  • Reproduced and modified state-of-the-art papers for in-house research project on spatial-temporal defect detection and presented to the whole group.
  • Created an AI-automated manufacturing-based labelling tool as a webapp and containerized with Docker for IPV.

Institute of High Performance Computing, A*STAR

11/1/2022 - 7/1/2024
A*STAR logo.

Research Intern (IHPC) (Nov 2022 - Jul 2024)

Institute for Infocomm Research, A*STAR

9/1/2022 - 1/1/2023
A*STAR logo.

Project Intern (I2R) (Sep 2022 - Jan 2023)

  • Worked in the Institute for Infocomm Research (I2R) Visual Intelligence division as a Student Researcher and Project Intern
  • Developed, tested and analysed two methods of detecting blurred images
  • Cleaned and modified datasets for testing and training purposes (e.g. gaussian blurring images, reformatting images)
  • Researched different edge operators and developed feature engineering paired SVM model for classification task
  • Tested various CNN architectures for optimized performance on this specific problem (e.g MobileNetV2)
  • Made detailed reports on developments in research every week and presented report to supervisor (worked with numbers and graphs)

Wanpo Tea Shop

3/1/2022 - 5/1/2022
Wanpo Logo

Barista

  • Assigned not only a barista position but a cook, cashier and also assigned to topping station
  • Responsible for preparation of drinks from mixing to toppings to cooking the toppings like boba and also tea. Also responsible for handling of customers as a cashier.
  • Responsible for keeping clear communications between every station to make sure stock of tea and toppings are kept up at all times
  • Assigned with two trainees to train in procedures for cooking of toppings and tea

ClassDo

11/1/2021 - 3/30/2022
ClassDo logo.

Intern

  • Supported the Customer Success Team in an emerging Ed-Tech startup
  • Worked on self-help articles on the official website for users
  • Helped with tech support, assisting users with difficulties using the website
  • Created product demonstration videos for users to learn about different functionalities of the website
  • Webscraped data from websites, preprocessed the data and exported it to an excel file