TL;DR - Store Resume data as JSON and write node scripts to generate dynamic versions like PDF and text. Source here
My website is geared towards recruiters and folks to get to know me professionally. One of the things everyone looks at is the Resume. I offer my Resume as a PDF or a Text file as downloads on this site. When a recruiter reaches out, I just redirect them to my website and they can choose the format that works for them.
Over the years, I’ve found myself updating my Resume and tweaking it’s format. The issue is that I have to do it twice - one for the PDF and another for the text. I’m a lazy developer and thought to myself - let’s automate this such that if there’s an update, I can only update it once and the PDF/text versions would be generated automatically. Another advantage with this approach is the ability to write a generator for a new format - like
.docx that I plan on doing soon.
JSON - single source of truth
Here’s why I decided to use JSON to store the data:
- Easy to read and understand
- Self-describing syntax. By looking at the structure of a JSON document, it is possible to interpret the context
- Stored as simple text and can be easily parsed in multiple environments
The source file can be found here.
Creating a PDF
After looking at the libraries available in the ecosystem, I decided to go with pdfkit as it was the easiest to get started and offered a simple enough syntax.
The source code can be found here
Essentially, here’s what is happening:
- Get resume data as a JSON object
- Create an empty PDF document with margins
- Set the output path
- Render the different sections of the Resume
Creating a text file
Writing this generator was a breeze as there is no third-party library required. All you need is to generate a template string and output it to a file using the fs module.
Auto-generation during deployment
Now that we have our node scripts that generate PDF and text versions of our resume, it’s time to write a command to do so dynamically in the CI environment when this website gets generated.
We could also check-in these files into the git repo by manual generation. However, this is a problem for me as I tend to forget things unless enforced by a rule. When merging a PR, it is possible that I forget to run the node scripts locally. Hence, I decided to do this part in CI.
I created a dedicated
package.json script to generate these files, but integrated it in the same
build command that generates the website. This way, I didn’t have to change any of my CI/CD scripts.
Can I use this to generate my own Resume?
Absolutely! As long as you are using Node.js, you should be able to copy the source JSON format, tweak it to add your own data/work experience and run the respective generator to generate the file.
And that’s it! Any time I make a change to the source JSON data, all formats are regenerated and deployed automatically.