Hefeng Zhang | Portfolio
Let design guide our intuition, and data shape our decisions.
Hefeng Zhang
I'm Hefeng Zhang. My work focuses on bridging human-centered design with data analysis. I am currently pursuing a master’s degree in Information Science @ Cornell University, and I earned my bachelor’s degree in Interior Design from Parsons School of Design.
I position myself at the intersection of product management and data analytics, using machine learning models to identify complex patterns and transform data into practical product solutions for real world applications.
In my spare time, I enjoy spending time with my cat and going on small hiking trips while catching the sunset.
Work Experience
Hanshow Technology
- Led A/B testing and user intent analysis using behavioral data for search interface iterations, increasing click-through rate (CTR) by 12% and supporting data-driven decisions.
- Analyzed user behavior and sales data using SQL, Excel, and Python to identify optimization opportunities across search and recommendation features.
- Collaborated with product, engineering, and business teams to convert data insights into model-driven feature improvements and measurable product outcomes.
- Built SQL/Python pipelines to track daily feature performance (CTR, conversion, retention), reducing manual analysis time by 30%.
Sanjin Capital
- Built and evaluated classification models on 1M+ high-frequency market data points to predict trading signals, improving signal precision and consistency across backtesting scenarios.
- Developed Python-based backtesting pipelines to simulate strategy performance, increasing Sharpe ratio by 15% while monitoring risk metrics including maximum drawdown and volatility.
- Processed and standardized large-scale financial datasets, reducing missing and noisy data by 20% and engineering time-series features to improve model accuracy and robustness
Skills
Projects
Developed an interactive data visualization web interface processing structured JSON data from 1.5K data centers, enabling exploration of global data infrastructure patterns and associated energy usage. (prototype under portfolio section)
Developed a machine learning model using 1M+ historical visa records to predict approval probability and identify key decision factors, leveraging Pandas, PyTorch, and Scikit-learn.
Developed an IoT smart planter using Random Forest and XGBoost algorithms to predict optimal watering schedules from real-time humidity and temperature sensor data for different plant types, achieving 90%+ prediction accuracy.
Education
Coursework: AI Chatbots, RAG, AI Agents, AI for Business Applications, Deep Learning, Database Systems
Coursework: Python: Data, Science & Design, Linear Algebra, Multidisciplinary Calculus, Intro to Machine Learning, Intro to Data, Statistics with SPSS
Design
click to scroll
Click to open the full portfolio PDF.
Click to watch the project walkthrough.
Click to open the Smart Planter project PDF.
Let's connect.
I'm open to jobs, collaborations, and coffee chats. Feel free to reach out through any of the channels below.
2026 Hefeng Zhang The website icon is a tribute to my little dog, Leo.