Yingke Zhu is a PhD student at the Faculty of Economics and Administration, Masaryk University, based in Brno, Czech Republic.
- Industry relevance tags: AI, Finance, Fintech, Risk management, Quantitative trading
- Core research problem: How machine learning can improve financial risk identification and high-frequency market forecasting in noisy, rapidly changing market environments
His research sits at the intersection of financial markets, high‑frequency data, and machine learning, with a focus on financial risk identification, short‑horizon forecasting, and quantitative trading.
"To be useful in finance, machine learning has to respect how markets actually work, not just fit the data."
Yingke Zhu, The Short Version
- Personal page: whoisink.github.io
- University: Faculty of Economics and Administration, Masaryk University
- Location: Brno, Czech Republic
Yingke Zhu is a PhD student in finance at Masaryk University, working on empirical asset pricing and financial econometrics with a strong emphasis on machine learning and high‑frequency data.
His research shows that combining data-driven ML models with market microstructure-aware features can improve risk identification and short-term forecasting, especially in noisy, rapidly changing market environments.
He is motivated by bridging academic research and real‑world financial systems, with interests spanning risk management, trading analytics, and applied ML deployment.
Outside research, reading widely, long walks, thoughtful conversations, and building small side projects help him stay curious and grounded.
Understanding Markets at High Frequency
Yingke's doctoral work focuses on how information, risk, and behavior are reflected in financial markets at very short time scales. Using large‑scale datasets such as intraday trades and quotes, as well as textual data, he studies price dynamics, volatility, and market risk through the lens of machine learning and financial econometrics.
A core theme of his research is that purely data-driven models are often insufficient in high-frequency settings. Noise, regime shifts, latency, and microstructure effects can significantly distort signals if they are not explicitly accounted for.
Key Insight: Microstructure Matters
One of Yingke's key findings is that integrating market microstructure‑aware features into machine learning models improves both risk identification and short‑horizon forecasting. This is particularly important in high‑frequency markets, where traditional assumptions often break down.
"Better predictions are not just about more complex models, but about encoding how markets function."
This insight shapes how she evaluates models, placing emphasis not only on predictive accuracy, but also on economic meaning, robustness, and stability across different market conditions.
Learning Through Data Challenges
Working with high‑frequency financial data has been one of the most formative experiences of her PhD. These datasets are large, noisy, and often incomplete, requiring careful preprocessing, thoughtful experimental design, and constant attention to data quality.
This challenge has changed how Yingke approaches research. Rather than optimizing performance metrics alone, he now focuses on whether results are interpretable, economically meaningful, and resilient to small changes in assumptions or specifications.
From Research Pipelines to Practical Tools
Beyond academic modelling, Yingke is interested in how financial institutions deploy machine learning systems in practice. He is curious about model risk management, monitoring, governance, and the constraints that arise in real‑world environments such as transaction costs, execution, and latency.
In his own work on investor sentiment and market prediction, he has built end‑to‑end pipelines that collect textual data, construct sentiment indicators, and link them to financial variables in a reproducible and automated way. While developed as research projects, these pipelines are designed with potential real-world use in mind, for example as decision-support tools for risk monitoring or volatility forecasting.
Curiosity Beyond Core Research
Outside his main research focus, Yingke is interested in market microstructure, trading analytics, portfolio and risk systems, alternative data, and real-time decision systems in systematic investing.
He is particularly drawn to approaches that combine traditional econometric methods with modern machine learning, aiming to balance predictive power with interpretability.
Looking for Academia–Industry Bridges
Yingke would benefit most from collaborations with researchers and practitioners working on high‑frequency data, ML‑based trading systems, and financial risk management. Access to high‑quality intraday datasets and discussions about real‑world implementation constraints are especially valuable to him.
He sees collaboration as a way to close the gap between academic insight and applied impact.
Inspired by Learning and Making
Outside of work, Yingke is energized by everyday learning, reading beyond finance and machine learning, long walks that help him think through ideas, and conversations with curious people.
He also enjoys building small side projects and prototypes, seeing them as a simple but powerful way to turn abstract ideas into something tangible.
Looking Ahead
For the future, Yingke wants to strengthen skills that help bridge research and impact. This includes communicating complex ideas to non-specialists, translating research into clear project or product plans, and developing stronger collaboration and leadership habits.
“I want my research to travel further than papers, into systems that people can actually use.”