I'm a PhD Candidate in Accounting and a former commercial banker. I like to draw on my professional background to develop research projects. I am primarily interested in financial archival research, particularly topics in media and disclosure, debt contracting, and financial regulation.

I have dedicated considerable time to learning about generative AI and integrating large language models (LLMs) and other AI technologies in my research and teaching workflows. I created chat-lab.ai to share this interest and help other academics use LLMs and AI tools in their research and teaching.

In addition to working with AI & LLMs, my research interests have required that I collect, maintain, and/or work with large, often unstructured, datasets. This has led me to learn about data storage and infrastructure, including cloud-based and self-hosted solutions. Because I can't help myself, I now host self-hosted apps and services from my personal homelab.

Below is a summary of some interesting datasets that I have worked with.

• ProQuest Text and Data Mining (TDM) Studio

  Full-text archive of various newspapers and magazines, scholarly and trade journals, electronic theses and dissertations, and other sources.

• Commercial Loan Officer Signors on debt contracts

  Manually collected dataset of debt contracts from material contract disclosures in SEC filings, matched to Dealscan loan information. Data from filings was collected using traditional textual analysis, while individual signors data was collected with large language models from the OpenAI API.

• OpenSecrets

  Political contributions data collected from Federal Election Commission filings, compiled into structured format useful for academic researchers (and other interested parties) by OpenSecrets.org.

• Twitter API v2 for Academic Research (deprecated)

  Access to the full archive of historical tweets and, with approval for academic research, users also benefitted from rather large monthly download limits. Sadly, this very useful API was deprecated following Twitter's private sale in 2022.

• Kalshi Prediction Market Data

  Historical order book and candlestick data for prediction market contracts collected via the Kalshi REST API (v2), using RSA-PSS/SHA-256 authenticated requests. The dataset covers event market series across five categories (Economics, Financials, Politics, Culture, and Science & Technology), with minute and hourly-resolution price data including open/high/low/close bid and ask quotes, transaction prices, volume, and open interest. Data also includes event metadata and market settlement criterion.