I'm Oliver — a Computer Science student at Appalachian State University graduating in May 2026, with a minor in Mathematics and a Data Science certificate. I focus on machine learning, deep learning, and natural language processing. Right now, I'm also working as an ML Engineering Lead at an early-stage startup, which has been one of the most challenging and rewarding experiences of my time in school.

What drives me is understanding how models actually learn — particularly through attention mechanisms — and how the decisions we make around architecture, training, and data shape performance, generalization, and interpretability. I enjoy working close to the math, moving between theory and implementation, and letting experimental results guide the next step.

What I'm Working On

At the startup, I lead a small team building a production ML pipeline that harmonizes product attributes across 125K+ items from heterogeneous data sources. We use semantic embedding matching with sentence-transformers and GPU-accelerated inference. I also architected a multi-format data ingestion layer that auto-detects and normalizes JSON schemas into a unified attribute registry. It's been a crash course in shipping ML under real product constraints, and I collaborate directly with the founder on a weekly basis to integrate ML outputs into the live product.

On the research side, my main personal project is LexiMind — a 272M-parameter multi-task transformer (FLAN-T5-base) that I built to investigate positive and negative transfer when jointly training for abstractive summarization, topic classification, and emotion detection across literary and academic text. A big focus has been on designing controlled ablations and understanding when multi-task learning actually helps versus hurts, rather than assuming joint training is always beneficial. I'm currently authoring a paper for submission based on this work.

Beyond that, I've built an F1 race outcome predictor using classical ML methods and feature engineering, and PlayAxis, a full-stack sports platform integrating multiple external APIs with a FastAPI backend and PostgreSQL. These projects sharpened my ability to reason about real-world datasets, evaluation metrics, and end-to-end system design.

Background & How I Learn

My coursework has covered Applied Machine Learning, Reinforcement Learning, Numerical Methods, Data Structures & Algorithms, and Database Systems. I also completed the Machine Learning Specialization by Andrew Ng, and I regularly read foundational and modern ML literature to stay grounded in both the classics and what's new.

I learn best by building. I implement models from scratch, validate ideas through experiments, and refine systems based on evidence rather than assumptions. NLP, representation learning, and applied deep learning are where I spend most of my time — and I genuinely enjoy explaining complex ideas clearly, whether through writing, diagrams, or code.

What I'm Looking For

I'm looking for ML Engineering or research-oriented Software Engineering roles where I can:

• Contribute to model development, experimentation, and production ML systems
• Work alongside experienced engineers and researchers who push me to grow
• Keep building a strong foundation in ML systems and applied research

This site is where I share my projects, technical work, and what I'm learning along the way. Thanks for stopping by.