Muhammad Talha Shafqat
I’m Muhammad Talha Shafqat, a Master’s student in Data Analytics at The Pennsylvania State University, with hands-on experience building machine learning and deep learning models for real-world problems.
About
I’m Muhammad Talha Shafqat, a Master’s student in Data Analytics at The Pennsylvania State University, with hands-on experience building machine learning and deep learning models for real-world problems. My work spans predictive modeling, statistical analysis, and data mining across genomic, clinical, and image datasets, with a focus on preprocessing, feature selection, validation, and interpretation.
I’ve developed end-to-end pipelines in Python and SQL to wrangle messy data, train and evaluate models using techniques like K-fold cross-validation and SHAP, and translate results into actionable insights. From identifying a 19-gene panel for hepatocellular carcinoma to building cancer and diabetes classifiers with strong accuracy, I enjoy turning data into trustworthy, interpretable outcomes.
- 1 year's experience
- Joined September 2026
Industry experience
- Healthcare
- Life Sciences
- Education
- Professional Services
More experts in Valley Stream, United States
Work experience· 4 roles
Leading a data-mining project analyzing approximately 30,000 observations and 13 variables from a BLS-based occupation dataset to study relationships among AI impact, automation risk, salary, education, experience, and job demand. Cleaned and prepared data in Python and SQL, including numeric conversion, missing-data handling, outlier treatment, and categorical encoding for regression, classification, and clustering workflows.
Trained CNN-based deep-learning models for skin cancer diagnosis using a Kaggle image dataset with Keras/TensorFlow. Combined model architectures to improve classification performance and evaluated results using accuracy, precision, recall, F1-score, and ROC-AUC against baseline single-model approaches.
Developed hybrid deep-learning models and a blending ensemble for diabetes prediction. Preprocessed a clinical diabetes dataset using robust scaling and class balancing, then applied K-fold cross-validation to evaluate performance. Used SHAP-based feature importance analysis to interpret model behavior and improve interpretability.
Education
Hire Muhammad Talha Shafqat today
Post a job and invite Muhammad, or contact them directly. No hidden commission.