Li Jiatong | Quantitative Finance Researcher

Footprint

Education

Work

News

Mar 2026

Website Upgrade

2026年3月12日网站更新以下功能：
1. 首页鼠标悬停在名字部分时，新增卡顿特效
2. Contact部分新增联系地址、微信与IG链接，点击可跳转
3. 新增留言功能，可直接发送信息至我的邮箱
4. 新增Footprint部分，地图可缩放或滑动，点击节点可展示具体经历
5. Research部分新增abstract与download功能

Feb 2026

Career News

I have just finished my research assistant work at Department of IEDA, the Hong Kong University of Science and Technology

Dec 2025

Academic News

I will be starting my PhD program at the Hong Kong University of Science and Technology this August, supervised by Prof. Jiang Wei

Contact

Phone

+852 6907 9560

jiatong-li-scofield

Address

HKUST, Clear Water Bay, Kowloon, Hong Kong

Instagram

@scofieldcys

WeChat

@Scofield

Feel free to reach out — whether it's about research collaboration, career opportunities, or just a quick hello. Drop me a message below and I'll get back to you as soon as possible.

Name

Subject

Message

Education

PhD in Industrial Engineering and Decision Analytics

The Hong Kong University of Science and Technology

Research Topic: Quantitative Finance
Supervisor: Prof. Jiang Wei

Aug 2026 – Present

MEng in Industrial Engineering and Operations Research

University of California, Berkeley

GPA: 3.96 / 4.0
Concentration: FinTech

Aug 2025 – May 2026

BSc in Investment Science and Finance Analytics

The Hong Kong Polytechnic University

GPA: 3.78 / 4.3 · First Class Honor
Dean's Honor List · People Charity Foundation Scholarship · URIS Scholarship · Applied Mathematics Scholarship

Aug 2021 – May 2025

Publication

Working Paper: Decentralized Mining in Tokenomics

Coauthor with Jiang Wei, Qin Cong and Xiong Xiong

Research

A Dynamic Mining Model for Bitcoin Economics

Research Assistant · Supervised by Prof. Jiang Wei

Jan 2026 – Mar 2026

Extending the dynamic mining model with multiple control variables and jump risk
Solving the HJB equation of the miner's optimization model numerically using the finite difference method
Calibrating the dynamic mining model based on Bitcoin address-level data

↓ Download Demo

On Optimal Consumption under Consumption Constraints

Individual Research · Supervised by Prof. Yu Xiang

Sep 2024 – Apr 2025

Extending the Merton's model and deriving the HJB equation for the optimal consumption problem
Deriving the explicit solution of the HJB equation using methods including duality transformations
Conducting numerical analysis to verify the robustness of the model

Deep Learning Detecting Financial Fraud via NLP

Individual Research · Supervised by Prof. Tian Feng

Sep 2023 – Aug 2024

Applied Chinese NLP and LDA on text data crawled from social media of listed companies based on Gensim, NLTK and spaCy
Fine-tuned Bert-Base-Chinese model for feature selection based on Transformers and Tokenizers
Constructed a Bi-GRU embedding pretrained model for financial fraud detection based on TensorFlow

Experience

Quantitative Research Intern

China Universal Asset Management · Shanghai

Department of Index and Quantitative Investment

Jun 2025 – Jul 2025

Developed dividend stock factors based on fundamental and macro analysis, through data processing and industry neutrality
Constructed index enhancement products based on Dividend Index through weight optimization and factor stock selection
Studied the use of ETFs by overseas institutional investors and replicate cuttingedge literature

OTC Derivatives Trading Intern

Shenwan Hongyuan Securities · Shanghai

Financial Innovation Headquarters

Sep 2024 – Jan 2025

Pricing OTC exotic stock options using the Monte Carlo simulation to assist the quotation for private funds
Solving high-dimensional PDEs in pricing models using deep neural networks and BSDEs
Using a single-step survival method to significantly reduce the variance when calculating Gamma

Quantitative Research Intern

SGD Private Fund Management · Shenzhen

Arbitrage Strategy Team

May 2024 – Aug 2024

Establishing a domestic and international gold statistical arbitrage strategy through ETFs and LOFs
Developing ETF premium/discount arbitrage strategies and optimizing execution to minimize market impact
Tested and validated CTA strategies using historical data to assess performance and ensure robustness

Skills

Coding

Python C++ Matlab R HTML

Tools

Git LaTeX Markdown

Others

Certificate

Finance

Programming

Language

IELTS 7.5 GRE 322

Using address-level blockchain data, we show that Bitcoin, as a virtual resource, does not conform to Hotelling's rule: miners' supply decisions respond to inventory levels but not to prices. To explain this, we develop a dynamic mining model that balances the option value of holding Bitcoin against its opportunity cost. In this model, partial irreversibility of supply ensures a positive option value, but high jump risk reduces it, leading miners to supply at rates largely independent of prices and proportional to inventories. Calibrated to 2014-2023 data, the model matches miners' inventory dynamics and the evolution of average transaction fee rates.

This paper extends Merton's classical optimal consumption-portfolio selection model by introducing consumption constraints that reflect realistic behavioral and institutional restrictions. We formulate the optimization problem and derive the corresponding Hamilton-Jacobi-Bellman (HJB) equation. Through duality transformations and advanced analytical techniques, we obtain explicit closed-form solutions for the value function and optimal consumption strategy. Numerical analysis is conducted to verify the robustness of the theoretical results under varying parameter regimes, demonstrating the impact of consumption constraints on investor welfare and portfolio allocation.

This research applies deep learning and natural language processing techniques to detect financial fraud in Chinese listed companies. We crawl and process text data from corporate social media accounts, applying Latent Dirichlet Allocation (LDA) topic modeling using Gensim, NLTK, and spaCy for initial feature extraction. A fine-tuned BERT-Base-Chinese model is employed for advanced feature selection, leveraging the Transformers and Tokenizers libraries. The core detection framework is built upon a Bi-directional Gated Recurrent Unit (Bi-GRU) architecture with pretrained embeddings implemented in TensorFlow, achieving significant improvements in fraud detection accuracy compared to traditional approaches.

Jiatong Li Scofield

Footprint

News

Contact

Education

Publication

Research

Experience

Skills

Certificate