01
Predictive machine learning
Build features and compare tree ensembles for forecasting, stock ranking and structured prediction tasks.
Python · scikit-learn · LightGBM
Explore related workUBC Statistics · Class of 2027
Ziyang Zhou
周梓洋
Machine learning, grounded in statistics.
I'm Ziyang Zhou (周梓洋), a UBC Statistics undergraduate focused on applied machine learning. My strengths span feature engineering, predictive modeling and rigorous validation, with work in forecasting, financial prediction and AI systems.
Ziyang Zhou
Applied Machine Learning · Statistical Modeling | UBC Statistics
01 / FOCUS
Where I focus
Three connected areas: building models, testing their reliability and turning them into useful applications.
02
Validation & statistical modeling
Design temporal and grouped validation, check information leakage and use statistical reasoning to interpret model behavior.
Cross-validation · Backtesting · GLM
Explore related work03
Applied AI systems
Explore computer vision and model unlearning, and build LLM tools with explicit workflows and inspectable outputs.
PyTorch · RetinaNet · LLM tools
Explore related work02 / SELECTED WORK
Selected ML & AI Projects
Machine Learning CompetitionJul 2026
CSI 300 Portfolio Modeling Challenge
A LightGBM competition project that ranks stocks for a five-stock allocation and compares models across historical time periods.
PythonPandasLightGBMScikit-learn
View projectComputer Vision CompetitionJul 2026
Neural Debris Removal Competition
A Kaggle project comparing ways to reduce unwanted detections in a supplied RetinaNet model while retaining useful predictions.
PythonPyTorchRetinaNetScikit-learn
View projectAI-Assisted Developer ToolJun - Jul 2026
ZZYAgent: Local Coding Assistant
A Python coding-assistant prototype using DeepSeek, with file tools, test execution, session history, and context management.
PythonDeepSeekTyperRich
View project03 / EXPERIENCE
Experience
Work Experience
Data Analyst Intern
Anhui Shusheng Data Technology Co., Ltd.
May 6-31 & Aug 3-31, 2026Hefei, China
- Collected, cleaned, and standardized electricity sales, load, weather, holiday, and customer data for weekly and monthly consumption forecasting.
- Used correlation analysis to narrow more than 20 candidate features to 12, and helped build and tune models including Random Forest.
- Contributed to weighted model ensembles, adjusting weights through historical backtesting and comparing them with individual models.
- Prepared forecast charts and comparisons for project reports and client presentations.
Work Experience
Actuarial Department Intern
Shanghai Guangbo Hi-intelligence Technology Co., Ltd.
Jun 8 - Jul 31, 2026Shanghai, China
- Used SAS to clean and standardize policy and business data, and reviewed monthly operating trends for actuarial analysis.
- Assisted with screening, validating, and calibrating more than 15 risk factors for GLM-based insurance pricing; compared model results across factor specifications.
- Independently extracted and checked rate-table factors, and co-authored a standard operating manual to make repeated checks more consistent.
- Selected and shared daily insurance and financial industry updates with the team.
04 / BACKGROUND
The statistical foundation
Machine learning is my focus. Statistics is the foundation. I bring them together through predictive modeling, disciplined validation and practical AI systems.
I'm Ziyang Zhou (周梓洋), a fourth-year Statistics undergraduate at the University of British Columbia (UBC), expecting to graduate in May 2027. My strengths are applied machine learning, statistical modeling and model evaluation. I work across the modeling process: defining prediction tasks, engineering features, training models and testing how well they generalize.
My core toolkit is Python, scikit-learn and gradient-boosted trees, including LightGBM. I apply these methods to forecasting and financial prediction, with particular attention to information leakage, temporal validation and the value of each new feature. In industry, I worked on electricity consumption forecasting at Shusheng Data and GLM-based insurance pricing support at Guangbo.
05 / TOOLKIT
Skills & Tools
Machine Learning
PythonScikit-learnLightGBMXGBoostCatBoostRandom ForestFeature EngineeringWeighted Ensembles
Model Evaluation & Forecasting
Cross-validationTemporal ValidationGroupKFoldBacktestingFeature SelectionTime-Series ForecastingGridSearchCV
Statistical Foundations
StatisticsProbabilityRegressionGLMSASRK-MeansDBSCAN
AI Project Experience
PyTorchRetinaNetModel UnlearningGloVeLDALLM Tool CallingContext ManagementLocal Qwen / Ollama
Experiment & Data Tooling
PandasNumPyJupyter NotebookMatplotlibPlotlyTushareGitPytest
Software & Applications
Data StructuresAlgorithmsJavaC++TypeScriptNext.jsReactStreamlitAI-Assisted Development
06 / CONNECT
Contact
I am looking for internships and graduate opportunities in machine learning, predictive modeling and applied AI. I expect to graduate from UBC Statistics in May 2027. For opportunities or technical collaborations, email is the best way to reach me.