Collaborating with Claude Code to Update My Academic Website

Today I had an interesting experience collaborating with Claude Code to completely overhaul my personal academic website. As a PhD student in Geological and Earth Sciences at UT Austin, I needed to update my GitHub Pages site with real professional information instead of the placeholder content that had been sitting there.

The Challenge

My website (ktwu01.github.io) was built using the Academic Pages template, but it was filled with sample data - generic publications, placeholder talks, and mock teaching experiences. I had all my real information organized in LaTeX files in my cloud storage, but manually converting everything to Jekyll markdown format would have been time-consuming and error-prone.

The Collaboration Process

What impressed me most about working with Claude Code was its systematic approach:

1. Analysis Phase: Claude Code first examined the existing website structure and my source files to understand the data format and required transformations.

2. Content Verification: It carefully cross-referenced my LaTeX CV sections with the information I provided, ensuring complete accuracy including proper author attributions (asterisks for corresponding authors) and citation formats.

3. Systematic Updates: The AI methodically updated each section:

   • Publications with proper DOI links and data availability
   • Conference talks and presentations with accurate venues and dates
   • Teaching experience with detailed responsibilities
   • Portfolio projects showcasing my AI/ML and geoscience work

Technical Details

The collaboration involved updating multiple Jekyll collections:

• _publications/ - Added my JGR: Space Physics paper and education research
• _talks/ - Included presentations from AGU, NASA workshops, and JSG symposium
• _teaching/ - Documented my TA experience with Earth in 2100 (520+ students)
• _portfolio/ - Showcased my AI projects including the NCAR-funded Noah-MP research

Quality Assurance

What really stood out was Claude Code’s attention to detail. When I asked it to double-check everything, it found and fixed several accuracy issues:

• Missing corresponding author asterisks in citations
• A forgotten AGU presentation that should have been included
• Proper venue name formatting

Lessons Learned

This experience highlighted how AI tools can be incredibly effective for:

• Data Transformation: Converting between different formats (LaTeX → Markdown)
• Accuracy Verification: Cross-referencing multiple data sources
• Best Practice Compliance: Ensuring proper Jekyll formatting and academic standards
• Systematic Organization: Managing complex, multi-file updates

The key was providing Claude Code with comprehensive source materials and being explicit about accuracy requirements. The “double-check everything” approach caught important details that could have been missed.

Looking Forward

As someone working at the intersection of AI and geoscience, this collaboration was a practical demonstration of how AI tools can augment academic workflows. The website now accurately reflects my research on explainable AI for Noah-MP land surface modeling, my internships developing AI evaluation systems, and my broader work in space physics and Earth system science.

The updated site serves as both a professional portfolio and a testament to effective human-AI collaboration in academic contexts.

This post was written on September 17, 2025, immediately after completing the website updates with Claude Code assistance.