Hello, I'm

Diing-Ruey
(Jimbo) Tzeng

Quantitative DeveloperML ResearcherIncoming UCLA MEng DS

I build systems where precision matters — from low-latency trading pipelines to graph-based ML models. Currently at VICI Holdings; heading to UCLA for MEng in Data Science, Fall 2026.

Scroll

About

At the intersection of
systems & intelligence

I'm a quantitative developer and ML researcher based in Taipei, Taiwan. I hold both a Master's and Bachelor's degree in Computer Science from National Tsing Hua University (NTHU), and I'm heading to UCLA in Fall 2026 for the MEng in Data Science.

Currently at VICI Holdings, Taiwan's leading proprietary HFT firm, I design and deploy automated futures and options strategies, and I've driven latency reductions of 50%+ in live execution systems.

My research background spans graph neural networks, contrastive learning, and recommendation systems — with a co-authored publication in Data Mining and Knowledge Discovery (DMKD, Springer).

Technical Skills

PythonC/C++SQLPyTorchGraph Neural NetworksPySparkLinuxKubernetesAWSGCPRecommendation SystemsLow-Latency Systems

3+

Years in industry

1

Publications

50%+

Latency reduced

1

Startups co-founded

Experience

Where I've worked

  • Designed and deployed automated futures and options strategies (maker & taker) coded into low-latency, scalable execution systems under risk constraints.
  • Uncovered hidden factors driving system performance, opening a new dimension in latency analysis and reshaping the company's understanding of execution pipelines.
  • Proposed and validated structural adjustments that reduced measurable latency by 50%+, prevented slippage, and improved execution quality of live trades.
  • Led cross-functional research initiatives with trading, infrastructure, and MIS teams.

Taipei, Taiwan

Education

UCLA

MEng, Data Science

Fall 2026Incoming

NTHU

MS, Computer Science

2020 – 2022GPA 3.8/4.3

NTHU

BS, Computer Science

2016 – 2020GPA 3.63/4.3

Research

Publications

Journal Article2024Graph LearningRecommendation Systems

Improving Graph-based Recommendation with Unraveled Graph Contrastive Learning

A novel approach to graph-based collaborative filtering that addresses representation collapse in contrastive learning frameworks, improving recommendation quality on benchmark datasets.

Data Mining and Knowledge Discovery (DMKD)·Springer, Vol. 38, pp. 2440–2465·Co-authored
ProjectGNNXGBoostBlockchain

Malicious Blockchain Entity Detection

NCHC-funded system achieving 89.1% F1-score in detecting malicious accounts and smart contracts on-chain.

Internship ResearchLightGCNCold-startPySpark

Graph-based Bank Recommendation

Graph recommendation framework for CTBC Bank mitigating cold-start using LightGCN and feature engineering on 20+ SQL tables.

Contact

Let's connect

Whether it's a research collaboration, quant opportunity, or just a good conversation — I'm happy to hear from you.

Diing-Ruey (Jimbo) Tzeng · Taipei, Taiwan · 2025