Research Profile
University of Sydney
MPhil in Computer Science
Research-based degree · 92/100
Supervisors
Josiah Poon and Matloob Khushi
Location
Sydney, Australia
Research Focus
Current Work
Cryptocurrency Price Forecasting
Under ReviewCryptocurrency Price Forecasting with Graph Neural Networks
Under ReviewKalman Filtering x MPC Stock Trading System
Under ReviewFinancial LLMs x Reinforcement Learning
OngoingHigh-Frequency Trading x Reinforcement Learning
OngoingNeuroAI
OngoingI am a Master of Philosophy student at the Department of Computer Science, University of Sydney, advised by Prof. Josiah Poon and Prof. Matloob Khushi. My research focuses on AI for Finance, with an emphasis on machine learning, time-series modelling, graph neural networks, reinforcement learning, and control-inspired decision systems for financial markets.
Before my research path, I worked in investment research and asset management, which gave me practical exposure to market dynamics, risk constraints, and data-driven financial decision-making. This industry experience now helps me connect AI methods with real-world financial problems.
I am currently a MicroMasters student in Statistics and Data Science at MIT.
Preprint · 2026
Yu Peng, Matloob Khushi, Josiah Poon
CryptoGAT introduces a lightweight graph attention network family for cryptocurrency forecasting, including a GAT baseline and an FGAT variant for feature-enhanced market modelling.
Three papers are currently under review at ICDM 2026 (CORE A*).
Launched AI Paper Reviewer, an open-source LLM-based academic paper review project.
Launched Tidewater Intelligence, a financial AI agent service.
One paper is currently under review at SIGKDD 2026.
Completed the WORLDQUANT University Applied Data Science Lab project.
Completed the JPMorgan Quantitative Research Experience Program.
MPhil in Computer Science
Advisor: Prof. Josiah Poon and Prof. Matloob Khushi
GPA: 92/100
MicroMasters in Statistics and Data Science
Advisor: Prof. John Tsitsiklis, Prof. Patrick Jaillet, Prof. Dimitri Bertsekas, Prof. Philippe Rigollet
Master of Information Technology in AI and Data Science
Grade: 82/100 (D); Computer Networks: 86 (HD); Deep Learning and Neural Networks: 78 (D)
Bachelor's Degree in Communication Engineering
Researcher / Investment Manager
Conducted investment research and portfolio analysis for private fund operations, translating macro, sector, and market data into risk-aware investment decisions. This work now informs my research on AI-driven forecasting and financial decision-making.
Produced macroeconomic and commodity-market research for institutional and individual clients, combining data collection, industry analysis, and concise research reporting for real-world financial decision support.
Completed eight end-to-end applied data science projects, covering data cleaning, ETL pipeline design, training-set preparation, supervised and unsupervised machine learning, and visualizations for explaining data characteristics and model predictions to non-technical audiences.
1. Completed a simulation focused on quantitative research methods
2. Analyzed a book of loans to estimate a customer's probability of default
3. Used dynamic programming to convert FICO scores into categorical data to predict defaults
The 6th Susquehanna Algorithmic Trading Competition (54/300)
Susquehanna & UNSW Fintech
The 2nd “Qianlong Cup” Private Equity Investment Competition: 6th overall and 2nd monthly among 4,000 products
CITIC Capital
The Third “ShenChengLunJian” Competition: Annual Outstanding Private Placement Product Award
XinHu Futures
Discovery Cup Internet Marketing Competition: Second Prize, Western Region
Ministry of Industry and Information Technology Education and Examination Center
District-Level All-Round Excellence Student in Tianjin (Top 1%, 3/270)
Tianjin Municipal Education Department