Hello, I'm
Kadeeja Zumreen
A data professional who believes the best insights come from equal parts technical depth and creative thinking, whether that's building ML models, designing dashboards, or finding the story hiding in a dataset.
Tools & Impact
Technical Stack
- Languages: Python, SQL, R, SAS, Java, Matlab
- Engineering & ETL: Alteryx, Excel (Power Query), Git, Linux/Bash, AWS SageMaker
- Visualization: Tableau, Power BI
- Libraries: Pandas, NumPy, Scikit-Learn, Matplotlib, Plotly, PyTorch
Machine Learning & AI
- Deep Learning: CNNs (VGG16), Transformers, Keras/TensorFlow
- Model Optimization: Quantization, Accuracy Recovery, QAT
- Computer Vision: Image Classification, Feature Extraction, PCA
- NLP: LLM Fine-tuning, VADER, LDA, Sentiment Analysis
- AI: Copilot, Cursor, LLM Models (GPT/Claude), Prompt Engineering for Research
Analytical Insights
- Advanced Statistical Modeling: Applied Regression, Statistical Inference, Probability, A/B Testing
- Predictive Frameworks: Time-Series Analysis, Demand Forecasting, Trend Modeling
- Data Storytelling & Logic: Experimental Design, ROI Modeling, Stakeholder Narrative Translation
Digital Trend Analysis
I architected a predictive framework to analyze how digital trends diffuse across YouTube, Reddit, and Google Trends. Using automated ETL pipelines to process 10 years of engagement data, I deployed Random Forest and NLP models to identify the specific sentiment and visibility signals that distinguish long-term cultural shifts from temporary engagement spikes.
Financial Market & Performance Analysis
Can we decode the daily fluctuations of a banking giant? By isolating the most critical drivers among 74 financial variables, I engineered a high-precision model that explains 98% of Citigroup’s price movements, turning market volatility into a structured data story.
Computer Vision & Age Detection
Estimating age from facial images is challenged by genetic and environmental variance. I improved classification accuracy by 19% through a data-centric redesign, strategically reframing 11 granular age classes into 5 demographic clusters. This project leveraged VGG16 Transfer Learning and PCA to optimize feature extraction across a dataset of 20,000 images.
"Kadeeja has been instrumental in helping us make sense of the data we collect around the world. She focused on analyzing usage statistics aand survey responses from our users, translating complex datasets into clear, impactful insights. Her interactive Tableau dashboards and visualizations provided stakeholders with a deeper understanding of SolarSPELL's effectiveness and reach. Furthermore, she has helped design and develop an improved data management system to streamline future analysis and reporting. Her dedication, attention to detail, and problem-solving mindset have significantly strengthened our data infrastructure. Kadeeja has also collaborated with other teams at SolarSPELL, providing data to support the needs of other interns. She quickly responds and provides data analysis on any area that is needed. She brings thoughtfulness, technical skill, and a genuine passion for impact to everything she does. Any team would be lucky to have her!"
Cassandra Barrett
Student Engagement Coordinator, SolarSPELL
