Projects

Selected case studies in AI content evaluation, natural language processing and technical communication — what the problem was, how I approached it, and what changed as a result.

2024

Amber-lit streams of generated text tokens flowing out of a dark terminal window, illustrating a fine-tuned GPT-2 model producing long-form text — Fine-tuned GPT-2 generating domain-specific long-form text, evaluated against the base checkpoint.

Text Generation Model with GPT-2

Senior capstone: fine-tuning GPT-2 for domain-specific writing.

PythonPyTorchHugging Face TransformersOpenAI API

2023 — Present

Grid of amber scoring bars across weighted rubric dimensions, representing a framework for ranking AI-generated content quality — The weighted rubric: factuality, instruction adherence, tone, structure and safety scored on calibrated bands.

AI Content Evaluation Framework

A rubric and workflow for grading model output at scale.

PythonSQLPandasLabel tooling

2022 — 2023

Stacked documentation pages linked by amber connector lines to reference, tutorial and API panels, showing a restructured docs library — Scattered pages restructured into a task-oriented library: quickstart, API reference and troubleshooting.

Technical Documentation Overhaul

Rebuilding developer docs for an AI product suite.

MarkdownGitLaTeXHTML/CSS

2022

Raw text messages flowing through four amber filter panels and emerging as clean labelled records, depicting an NLP annotation pipeline — Raw text moves through ingestion, pre-labelling, human review and validated export before reaching training.

NLP Data Annotation Pipeline

Turning raw text into clean, labelled training data.

Pythonscikit-learnSQLAWS