Arine¶
Technical Writer, April 2025 - Present
| Criteria | Description |
|---|---|
| Documentation | Document360, Confluence, SnagIt, Claude, Gemini |
| Subject matter/industry | Healthcare, software, UI |
| Points of contact | Knowledge Base team, product managers |
Currently I serve as a solo technical writer at Arine (uh-REEN), a healthcare optimization platform for for pharmacists, clinicians, and other medical professionals. Our documentation is considered proprietary and therefore I'm unable to share any direct writing samples. However, I can still freely discuss my professional work more broadly.
Key Projects and Responsibilities¶
Migration to Document360¶
Moving Arine's documentation out of static PDFs into a living knowledge base was not an easy accomplishment. After much window shopping, the Knowledge Management team selected Document360 as a content management system to handle the total turnaround for product docs. Our documentation is closed-source, which means coordination with our full-stack engineering team was necessary to build out our user authentication to ensure only clients could access the content.
After higher-level logistics had been determined, the sheer bulk of the work required removal of any unique or advanced style elements, then converting the content into a basic, readable format. CSS was finally applied across the entire knowledge base, reducing the manual work that had been going into formatting PDFs. Overall, the project accomplished a number of feats previously missing or in need of assistance within the Knowledge Management team:
- The KM team could now own the publishing process itself (previously a full stack responsibility), allowing documentation to be drafted and published in a CI/CD pipeline instead of only during major software releases.
- Total turnaround time (time from drafting content, SME review, and publishing) was reduced by over half as a result of the manual/tedious processes removed completely.
- Greater insight into user activity was now available, as this was previously a black hole due to our PDFs not tracking these metrics.
Writing and Editing AI Agent¶
AI adoption is a skill quickly becoming in-demand. I found difficulties at first integrating AI tooling into our build pipeline, including day-to-day work. Nevertheless, I took initiative within my team to identify use cases where an AI writing agent could provide and additional layer of coherence across our content. Setting up a basic Claude Code environment to work, I provided the following materials as my base context:
- A
claude.mdfile providing both a persona and workflow approach as the basis for my writing agent (Wricardo, I adoringly call him). - All of our user-facing documentation, converted into Markdown (using Pandoc) for easy ingestion.
- Our company style guide that emphasizes key vocabulary, formatting, and other language guidelines.
With this, Wricardo was able to read through our existing documentation, and perform the following tasks:
- Find existing content that did not conform to our style guide
- Determine areas of content which would need rewrites based on proposed new feature functionality (read from Jira tickets)
- Identify bullet points and table cells which were excessively wordy, indicating content could be reformatted for readability
Due to the smaller scale of content ingested by Wricardo, and formatting the content for AI readability, I avoided hallucinations and other contextual errors getting Wricardo to do as instructed.