Chun Chet Ng

In 2018, I joined Prof. Chan Chee Seng's CISIP lab as a research assistant and started my Ph.D. in September of the same year, working on computer vision problems in scene text spotting and low-light image enhancement. Since transitioning to industry, my work has focused on production AI systems for fintech: covering document intelligence, financial retrieval (RAG), and alternative data for MSME credit scoring.

Below is a list of my publications over the years:
- Conference
  - ICIP 2023 - Rethinking Long-Tailed Visual Recognition with Dynamic Probability Smoothing and Frequency Weighted Focusing [paper] [code]
  - ICPR 2022 - Extremely Low-Light Image Enhancement with Scene Text Restoration [paper] [code]
  - ICDAR 2021 - Competition on Integrated Circuit Text Spotting and Aesthetic Assessment [paper] [dataset]
  - CVPR 2020 - On the General Value of Evidence, and Bilingual Scene-Text Visual Question Answering [paper] [dataset]
  - ICDAR 2019 - Robust Reading Challenge on Arbitrary-Shaped Text - RRC-ArT [paper] [dataset]
  - ICDAR 2019 - Competition on Large-Scale Street View Text with Partial Labeling - RRC-LSVT [paper] [dataset]
  - ACMMM 2018 - Unprecedented Usage of Pre-trained CNNs on Beauty Product [paper]
- Journal
  - Signal Processing: Image Communication 2025 - Text In the Dark: Extremely Low-Light Text Image Enhancement [paper] [code]
  - Pattern Recognition 2024 - When IC Meets Text: Towards a Rich Annotated Integrated Circuit Text Dataset [paper] [code]
- Workshop
  - PRISM: Prompt-Refined In-Context System Modelling for Financial Retrieval [arxiv] [code]
    - Scalable Learning and Optimization for Efficient Multimodal AI Agents @ ICML 2026
    - Workshop on Logical Reasoning of Large Language Models @ ICLR 2026
    - The 2nd Workshop on Advances in Financial AI: Towards Agentic and Responsible Systems @ ICLR 2026
    - International Symposium on Large Language Models for Financial Services (FinLLM@IJCAI 2026)
  - AI-BAAM: AI-Driven Bank Statement Analytics as Alternative Data for Malaysian MSME Credit Scoring [arxiv]
    - Workshop on Logical Reasoning of Large Language Models @ ICLR 2026
    - The 2nd Workshop on Advances in Financial AI: Towards Agentic and Responsible Systems @ ICLR 2026
    - Workshop on Multi-Agent Learning and Its Opportunities in the Era of Generative AI @ ICLR 2026
    - Agentic AI in the Wild: From Hallucinations to Reliable Autonomy @ ICLR 2026
  - Cash Flow Underwriting with Bank Transaction Data: Advancing MSME Financial Inclusion in Malaysia [arxiv]
    - Agentic AI in Financial Services @ AAAI 2026
    - AI for Financial Inclusion, Risk Modeling & Resilience in Emerging Markets @ ACM ICAIF 2025

Collaborations with the following people are key reasons behind successful research projects:

Postgraduate Researchers:

Chee-Kheng Chng - University of Adelaide
Che-Tsung Lin - Chalmers University of Technology
Xinyu Wang - University of Adelaide
Jia Yu Lim - National University of Singapore

Undergraduate Students I Worked With:

Wan Jun Nah - University of Malaya
Mei Yih Tan - University of Malaya

My recent working experience:

Apr. 2024 - Present

AI Lens Group

Chief AI Officer.

Designed and delivered AI-powered credit scoring and document intelligence platforms supporting bank statements, payslips, ICs, invoices, financial statements, and bureau reports, achieving >95% extraction accuracy at <5s per page on on-premise infrastructure.
Built end-to-end pipelines for automated document extraction, validation, analysis, credit scoring and report generation using FastAPI, Python, React, PostgreSQL, Docker, RabbitMQ, and Azure.
Provisioned and maintained Azure production infrastructure and cloud resources.
Set up CI/CD pipelines with GitHub Actions and self-hosted runners. Implemented branch gating, containerised deployments, and a structured Release Readiness Checklist (Azure/GCP review, DB migration validation, rollback planning) across production releases.
Helped establish sprint structure (8:2 feature-to-reliability ratio, 20% test coverage target), ClickUp task governance, and weekly engineering sync; supported developers on backend, DevOps, OCR, and AI system design.
Contributed to an R&D track with workshop publications at ICAIF, AAAI, ICLR, and ICML, translated research into production applications, and helped the team secure 3rd place at the ACM ICAIF 2025 Agentic Retrieval for Financial Documents (AI Grand Challenge on FinAgentBench) on Kaggle.

Oct. 2023 - Mar. 2024

Etch Media

Vice President of Engineering.

Integrated an LLM-based classification pipeline to map CV-detected object and person labels to contextual advertisement categories, enabling real-time targeted ad delivery.
Engineered a C# media playback service for custom LED display boards, handling advertisement scheduling and real-time content switching.
Designed and integrated backend microservices bridging AI inference engines for multi-attribute vehicle recognition (make, model, colour) and facial analysis (gender, age, emotion) into production ad delivery workflows.
Set up CI/CD pipelines with GitHub Actions and containerised deployments; improved web security via Cloudflare and maintained Azure cloud resources.

2023 - 2025

Tibame Image Recognition

Instructor.

Taught hands-on OCR classes covering license plate recognition and signboard text spotting using PaddleOCR and MMOCR. [video]
Course materials open-sourced: License Plate Recognition, Signboard OCR.

Jan. 2023 - Sept. 2023

AkiraKan

Part-time ML/AI Research Scientist (3D & AI).

Handled OCR-related projects including lamp post ID extraction and license plate recognition based on traffic scene images.

Mar. 2020 - Dec. 2022

ViTrox Corporation

Research Intern.

Worked on industrial OCR methods for detecting and recognizing defective characters on chip images under low contrast, blur, broken-stroke, and other aesthetic-defect conditions.
Co-organized ICDAR 2021 RRC - ICText with Baidu and partner universities, contributing to the largest Integrated Text on Chips Dataset (ICText).
Proposed a curriculum-learning training strategy to improve character recognition performance on ICText; produced two publications from the research project.
Supported undergraduate interns on image classification and object detection projects.

Sept. 2018 - Dec. 2023

University of Malaya

Research Assistant.

Researched scene text detection, recognition, spotting, and low-light image enhancement methods.
Contributed to benchmark datasets including ArT, EST-VQA, and ICText.
Ph.D. thesis: Optical Character Recognition Under Challenging Lighting Conditions.

Feb 2017 - Aug 2017

Theta Service Partner

Intern Software Engineer.

I joined the Centre of Excellence department to work on a bank loan software system. I was mainly responsible for bug fixing, implementing customers' requirements, and writing test cases.

References available upon request.

Alongside research and work, I contribute to the research community and society through the following volunteer work:

2019 - Present

Dhamma Earth

IT Volunteer.

Set up the organization's main website.
Developed an internal web application to track allowable requisites for Buddhist monks and nuns.

2017 - Present

Dharma Drum Mountain

Software Engineer.

Completed an internal E-receipt system to record received donations and payments in a systematic format.
Set up and migrated the legacy email systems to GSuite for the NGO.
Created an e-commerce website to sell books online and accept online donations.
Provide periodic updates and maintenance for the above projects.

I have also reviewed papers for the following conferences and journals: