ABScribe: Early PhD Backend Work

Hey friends,

During my early PhD work, I had the chance to contribute to the backend development for ABScribe, a research project exploring how writers can rapidly explore text variations using large language models.

What I Built

I worked on the Python backend infrastructure, which gave me hands-on experience with:

• Flask for API development
• MongoDB with MongoEngine ODM for data modeling
• Database design for managing document versions and user feedback
• Research software development patterns

The system handled the core data models (documents, chunks, versions, recipes) and provided APIs for the React frontend to interact with.

The Learning Experience

This project was my introduction to building software for academic research, which has different requirements than commercial development. I learned about designing flexible systems that can evolve with research questions and the importance of data collection for user studies.

Working with an interdisciplinary team of HCI researchers and frontend developers taught me how technical decisions impact research outcomes.

The Results

The project was published at CHI 2024 and showed significant improvements in writing task workload and user perceptions of the revision process. The backend systems I built enabled the research team to conduct user studies that wouldn’t have been possible otherwise.

You can find the project on GitHub if you’re interested in the technical details.

It was a great early PhD experience that combined my interest in software engineering with academic research.

Cheers,

Nathan