Modeling the Future: A Graduate Journey Through Machine Learning at UVA

I completed my first machine learning exam as a graduate student at the University of Virginia School of Data Science in November 2025. The exam focused on supervised learning and marked my second major assessment of the semester. The first was in Computation for Data Science, a course centered on data structures and algorithms that challenged me to think like a computer scientist and write efficient, scalable code. 

For me, the machine learning exam wasn't just another test. It felt like validation of how far I’ve come — from studying plant biology as an undergraduate to building a foundation in data science and machine learning at the graduate level.

What Machine Learning I Teaches

Machine Learning I is more than a coding or statistics class; it’s a deep exploration into how data becomes knowledge. The course provides a comprehensive introduction to predictive modeling, a cornerstone of both data science and artificial intelligence. It blends theory with practice, guiding students through both the mathematics and the mindset of model building.

Throughout the semester, we explored topics such as:

• Supervised Learning: linear regression, logistic regression, K-nearest neighbors, decision trees, and neural networks (multilayer perceptrons)
• Unsupervised Learning: K-means clustering, principal component analysis (PCA), and dimensionality reduction
• Model Evaluation: accuracy, RMSE, AUC, and cross-validation—along with the crucial trade-offs between bias, variance, and interpretability

By the end of the course, students are expected to do the following:

• Explain the distinction between supervised and unsupervised learning and their role in data science
• Apply and interpret linear and generalized linear models on real-world datasets
• Implement flexible methods such as KNN and decision trees and evaluate their predictive performance
• Use clustering and PCA to uncover hidden patterns in data
• Develop and compare neural network models to traditional approaches
• Evaluate models rigorously and communicate results effectively to both technical and non-technical audiences

Machine Learning I doesn’t just teach algorithms — it trains us to think critically about assumptions, interpret results meaningfully, and communicate insights with clarity.

A Personal Turning Point  

Coming from a background in plant biology, this course represents a personal evolution. Each lecture and assignment pushed me to think more mathematically and computationally, connecting ideas across statistics, optimization, and programming.

Every model I built — whether linear, multiple, or logistic regression; multi-class or Poisson regression; ridge, lasso, or elastic net — reminded me that growth in data science is iterative, much like the training process itself.

Machine Learning I helped me see data not just as numbers, but as patterns, processes, and possibilities waiting to be modeled.

Looking Forward: Where Curiosity Leads Next

As my master's journey unfolds, I’m always looking toward the next chapter. This semester, I’ll take electives outside the data science program, and a few courses stand out:

Geometry of Data (Computer Science Department) 

This course explores the mathematical foundations of high-dimensional geometry and manifolds, the very structures that underpin modern AI. It connects directly to my growing interest in physics-informed neural networks (PINNs).

Physics-Informed Neural Networks 

PINNs blend deep learning with physical laws to solve differential equations, enabling models that are both data-driven and physically consistent. I hope to study this field at the Ph.D. level, as it represents a compelling intersection of mathematics, physics, and machine learning — where my background in physics and my passion for computation converge. 

Quantum Computing and Quantum Machine Learning 

I also plan to take a quantum-focused elective, either Quantum Machine Learning or an independent study in quantum computing. This emerging field represents the frontier of computation, where data science meets quantum mechanics to redefine what’s computationally possible.

Closing Thoughts 

The Machine Learning I exam in November wasn’t just an academic checkpoint, it was a milestone in my ongoing transformation. It symbolized how I’ve moved from studying biological systems to modeling them, from observing phenomena to predicting them, and from curiosity to capability. 

At the UVA School of Data Science, I’m encouraged to pursue my deepest curiosities, and I plan to continue doing exactly that – one model, one dataset, and one quantum leap at a time.

Author

Shiraz Robinson II

MSDS Residential Ambassador