Like most people, I don’t update my resume nearly as often as I should. And when I do, I usually rely on my memory to piece together the details of past jobs, projects, and technologies I’ve worked with. But this time, I wanted to take a different approach—one that leveraged AI to help me create a “maximal” resume, something that captured everything before tailoring it down for specific job applications.
Step 1: Mining My Work History with NotebookLM
A few months ago, I had fed NotebookLM a bunch of my labor logs from my time at Zedasoft. My goal back then was to rediscover forgotten tasks and technologies I had worked with—things that might not have stuck in my memory but were worth remembering. That experiment worked well, so I decided to expand the idea.
If I wanted a truly complete resume, I’d need to pull in more sources:
- My Wabtec labor logs
- Git logs from my Oway and Tarterware projects
With all of this, I could let NotebookLM generate the first draft of my maximal resume. After several iterations of back-and-forth with the AI, I had something that was surprisingly accurate in content but terrible in formatting—chunky blocks of text, weirdly ordered sections, and nothing that looked remotely like a polished resume.
Step 2: Cleaning Up with ChatGPT
I needed something that could take this messy but content-rich document and structure it properly. Enter ChatGPT-4o-mini-hi.
I gave it a straightforward but specific prompt:
“Please act in the role of an HR expert and excellent writer and editor. Using the downloaded PDF as a guide, please create a maximal resume in the style of the PDF that combines information from the original with the additional information that follows.”
Then, I fed it:
- The last best version of the NotebookLM-generated resume
- My existing resume, which had a formatting style I liked
ChatGPT processed the request for nearly two minutes (which, let’s be honest, feels like an eternity in AI time). But when it finished, it nailed the structure. The formatting was on point, and the content was well-organized. Of course, there were still some issues—like mistakenly categorizing GCP as an operating system—but these were easy enough to correct.
Step 3: Final Human Touch
With the AI-generated draft in hand, I pasted it into Google Docs and did a final pass. This involved:
- Smoothing out phrasings to sound more natural
- Fixing inaccuracies (sorry, I wasn’t actually a “principal developer”)
- Removing any AI-induced exaggerations
At this point, I had a strong, AI-assisted resume that was better than anything I could have written from scratch in the same amount of time.
Key Takeaways
- AI tools like NotebookLM are great for extracting information from scattered sources, but they struggle with formatting.
- ChatGPT 3o-mini-hi was excellent at structuring and rewriting, but it still needed a human editor to fine-tune the details.
- The best results came from combining multiple AI tools, rather than relying on just one.
Final Thoughts
This experiment saved me hours of work and gave me a maximal resume I can now tweak for specific roles. AI didn’t replace my judgment, but it made the process significantly faster and more efficient.
If you’ve been dreading updating your resume, consider giving AI a shot—it might surprise you!