Marcus Lee
Data Scientist
Pittsburgh, PA 路 (412) 555-0193 路 marcus.lee@example.com 路 github.com/example
Summary
Data scientist with a year of industry data science after an M.S. in data science and a B.S. in mathematics and computer science. Builds machine learning and predictive modeling in Python, scikit-learn and SQL, from exploratory data analysis to model development on AWS, and explains results with clear data visualization. Recent work in natural language processing and applied artificial intelligence.
Education
M.S. in Data Science, Allegheny Institute of Technology, 2025
B.S. in Mathematics and Computer Science (algorithms, quantitative methods, data modeling, databases), Allegheny Institute of Technology, 2023
Skills
- Machine learning
- scikit-learn, TensorFlow, PyTorch, Deep learning, NLP
- Programming languages
- Python (pandas, NumPy, Matplotlib), SQL, Java, R
- Analysis
- Statistical analysis, Exploratory data analysis, Tableau, Power BI, Spark
Experience
Data Scientist I, Ridgeline Freight
Pittsburgh, PA 路 Aug 2025 to present
- Built a machine learning model in Python and scikit-learn that predicts late deliveries 48 hours ahead with 81% precision, now used by 30 dispatchers.
- Ran exploratory data analysis on 3 million shipments with SQL queries and pandas, and presented the findings to operations with Tableau dashboards.
- Engineered a data pipeline with 25 data quality tests for model features on the AWS cloud platform, working with the software engineering team.
- Applied natural language processing to 200,000 driver notes, a first analysis of that unstructured data, and flagged the 4 top causes of delays.
- Used clustering and route optimization to redesign 40 delivery routes, cutting miles driven by 9%.
Data Science Intern, Keystone Health Analytics
Philadelphia, PA 路 Summer 2025
- Trained a TensorFlow deep learning model for flagging suspect insurance claims that caught 12% more cases than the rules-based system.
- Created Power BI dashboards on model performance for the claims team.
- Helped move feature preparation from notebooks to Spark.
- Helped build a forecasting model for patient volume with data mining on past records.
Thesis, Computer Vision for Retail Shelves, Allegheny Institute of Technology
Pittsburgh, PA 路 Sep 2024 to May 2025
- Trained a PyTorch computer vision model on 40,000 images that detects out-of-stock shelves with 92% accuracy.
- Compared model architectures with statistical modeling, probability and hypothesis testing, and published the code on GitHub.
- Helped present the results to retail partners with a store operations case.