Gurleen Kaur | Data Scientist — ML & GenAI
Machine Learning, GenAI & Data Applications
I turn complex, real-world data into useful models and tools.
About Me
I build GenAI workflows that identify signals in unstructured communications, geospatial platforms that bring multiple data sources into one view, and predictive models using behavioral and sensor data. My work spans aerospace, gaming, industrial analytics, and biomedical research. Across these domains, I enjoy figuring out what the data can tell us and building something useful from it.
Featured Work
Radio-Frequency Interference Monitoring
A Databricks-hosted prototype integrating five data sources, including ADS-B, an existing spoofing API, and aircraft event data, into a unified 2D geospatial view using H3 hexagons. Coverage includes North America, Europe, and selected Asian countries. Currently in alpha development.
GenAI Order-Signal Detection
A GPT-based workflow that scores customer-supplier communications for potential future orders and generates alerts for human review. Correctly identified orders represented approximately $55K in parts value.
Forecasting Sleep-Related Respiratory Events
CNN and LSTM forecasting models in PyTorch using multichannel EEG and physiological signals. Processed and aligned recordings from 470+ patients with respiratory-event annotations. Used subject-wise cross-validation, multiple prediction horizons, and explainable AI techniques.
Professional Experience
Data Science Intern, Boeing — Jan 2026 to Present, Vancouver, BC
- Built GPT-4o-based workflows in Databricks and MLflow to extract info from unstructured maintenance text
- Combined multi-source data to prepare features for inventory optimization and demand forecasting
- Created analyses, visualizations, and presentations for customer-facing aviation projects
Undergraduate Researcher, MIAL Lab, SFU — May 2025 to Present, Burnaby, BC
- Built CNN-based deep learning models for EEG-based hypopnea event forecasting
- Implemented subject-wise cross-validation to prevent data leakage and ensure generalization
- Applied uncertainty estimation and abstention for reliable high-confidence predictions
- Used SHAP to interpret model predictions and analyze feature importance
Data Science Intern, Kabam Games — Sep 2025 to Dec 2025, Vancouver, BC
- Built Tableau dashboards tracking null rates, data drift, precision, recall, and F1-score across AI models
- Built a CatBoost model to predict whether a player's total spending would exceed $100 within their first three months, using seven weeks of player data
- Approximately 1% of players were high spenders; evaluated the model using precision and recall
- Performed error analysis on false positives/negatives and identified model blind spots
- Conducted impact analysis to identify feature gaps and inform potential retraining decisions
Data Analyst Co-op, MineSense Technologies — Sep 2023 to Apr 2024, Vancouver, BC
- Detected sensor anomalies using PCA and feature engineering to predict failures
- Optimized SQL queries with dynamic and parallel processing, cutting execution time by 25%
- Developed automated Python test cases with cross-functional teams
- Validated GNSS datasets using statistical methods
Service Desk Technician, SFU — Sep 2023 to Aug 2025, Burnaby, BC
- Provided front-line technical support to faculty, staff, and students
- Assisted with MFA setup, remote server access, and OS deployments
- Logged and tracked requests in TeamDynamix and Confluence
LinkedIn |
GitHub |
Resume